The Ambulatory ICU
The concept of an Ambulatory Intensive Care Unit (A-ICU) developed because in most companies a small number of employees are responsible for a majority of the healthcare costs. The 80/20 rule says that 20% of the employees drive 80% of the cost. These employees generally are people with chronic diseases such as coronary artery disease and diabetes. The A-ICU is a different way of managing these “walking sick.” A clinician working on this design described the A-ICU as “boot camp” for chronic disease patients. The most difficult to control patients fit into this category and need “a boot camp” to get back on track and better control their chronic condition. ThedaCare, an integrated health system in northeast Wisconsin, recently worked with an employer coalition to design a completely different way for companies to care for their high-risk employees.
A business coalition of more than 100,000 members approached ThedaCare to develop a new model of care in their market. They attended a site visit at ThedaCare and were impressed with the organization's ability to redesign care processes with an emphasis of adding value (quality/cost) to the customer. The coalition came to ThedaCare after seeing it had the lowest charges in the state on public web sites (www.wicheckpoint.org) and reading a story in the Milwaukee Journal Sentinel about the ways ThedaCare was improving its quality using the ThedaCare Improvement System, which is derived from the Toyota Production System and tailored for healthcare.
The first series of meetings was devoted to understanding the employers' problem. There were two major problems, the first being costs were 40% higher in their market than the region average and the second was that no provider system was willing to work directly with them to improve this performance. The coalition's major cost drivers were employees with coronary artery disease with the co-morbidity of diabetes, musculoskeletal disease and cancer. The data indicated employees with coronary disease spend $93 million annually on health costs and was therefore the biggest opportunity for improvement.
Once the two parties agreed on the focus, the next step was to spend three days on value streaming the current state and then the new possible future state.
(Fig.1). Example of a care delivery value stream
ThedaCare required the customer to participate in this value stream so both parties would be completely in agreement as to what the goals were and how success would be determined. The purpose of value stream mapping is to clearly understand the set of steps necessary in delivering a product to a customer. The value stream is always focused on the customer with the intent to provide the least wasteful way of delivering a service. ThedaCare had never done a value stream for a new product with an employer customer before. It was unclear if this would work, especially since the stakeholders on the employer side had never participated in this process. It was critical to choose teams representing the multitude of stakeholders in this process. The team from the business coalition included human resource specialists, healthcare providers from the companies, the business coalition director and project manager. In addition the third party administrator data analysts were included as part of the coalition's team. Finally, an internist and emergency room doctor from the existing market also agreed to participate. The ThedaCare team was represented by a business development person, pharmacist, lean facilitator, two physicians, a project manager and an electronic health records specialist. Since the coalition had very little experience with lean or the Toyota Production System, the sensei — otherwise known as teacher — designed a three-hour education component to get the group started. In this case, the group studied the set of processes that would deliver improved clinical care to a group of patients with coronary disease and diabetes. This was chosen as a focus since many of the coronary patients had Type 2 diabetes as their underlying diagnosis.
The first day of the value stream documented the present situation in the market. This map pointed out there was tremendous fragmentation in the market with little electronic connectivity. There was poor communication between providers and multiple studies commissioned by the business group had shown significant over-utilization in the commercial population. These studies were supported by data from the Dartmouth Atlas as well, which showed this market to be the least efficient market in the state in resource utilization in the last two years of life of Medicare beneficiaries. Clearly, there was a significant opportunity for improvement and although it was difficult to quantify, 20% to 40% of $93 million being spent every year was in the ballpark of the potential.
In the second day of this exercise, the group began to deeply understand the patient population. There were approximately 10,000 patients with these diagnoses. A large picture of a funnel was created to understand how these patients might be identified for intervention. There were basically five ways of identifying the population. The first was through claims using medication records. This would allow us to identify the diabetics, which could then lead us to the patients with coronary disease. The medication claims might also help us identify the employees with coronary disease by determining if they were on cholesterol lowering drugs. A second way to identify this group was through lab data. The insurer who acted as the TPA for this coalition had developed a tool to identify chronic disease patients who were not meeting the goal of blood sugar or cholesterol control. A third mechanism was the use of HRA (Health Risk Assessments). Many of the employers were already having employees complete an HRA so that would be one way to obtain the information. A fourth way of determining the patient population was simply by having the patients identify themselves. This would require significant work on the employer's part in terms of educating and communicating with their employee population. Finally, the patient's primary care physician could identify the patients needing intervention if they had confidence in the new process.
Each one of these avenues for identification would require a dedicated process with a specific owner for it to happen. This would be part of the implementation infrastructure required. The other infrastructure would involve the actual clinical process itself. The mapping of the clinical process required the help of a number of clinicians. These people came half from the marketplace in question and half from ThedaCare. The group mapped out how this ambulatory care process would be different than what existed in the current market with s focus on eliminating the issues that had caused such inefficiency. In addition to understanding how to identify these patients, the group also needed to set some criteria as to who really needed to be seen in this intensive clinic. The clinicians set very specific quality goals. These included a blood pressure measure of 140/80 or above, A Hgb A-1C of greater than 8.0 and an LDL cholesterol of greater than 130 mg/dl. It was decided that these were the patients most likely to be very costly over the next two years and where an A-ICU could make the biggest difference in total cost avoidance.
As noted above, the patients would be identified by several different mechanisms. But once identified would need to be educated as to the fact the A-ICU existed and why it would be of benefit to them. No solutions were agreed to, but examples of financial incentives that were discussed included eliminating any co-pay the employee paid. Free supplies for diabetic management and no drug co-pays were considered. The goal was to have no barriers for patients to access this care.
Once recruited, the patient would require a different experience for the A-ICU to be successful. The map of the experience was completed on the third day. This was completely different than the initial state. Unlike a usual clinic where the doctor is paid on the number of patients seen and procedures completed, in this case the doctor would be salaried and the incentives would be based on health outcome improvement and total spend reduction. The clinic providers would be focused on doing the right things rather than just doing things. It was clear that a lot of focus needed to be placed on education. Stealing concepts from ThedaCare's experience with the redesign of inpatient care a pharmacist, diabetic dietician, nurse educator and social worker would all work as a team to improve each employees' health. In fact, the first visit would be designed to include the whole team who would develop a single care plan for the patient. This plan would drive future visits and any needs the patient might have. Providing 24-hour access to medical personnel was another basic component of the care since that would diminish unnecessary emergency room use. This part of the plan could be accomplished by the call center ThedaCare had already established. The call center, which is called ThedaCare On Call, uses evidence-based guidelines to advise patients 24 hours a day and getting the patient to an appropriate caregiver if necessary. The design also included a “patient advocate” who could be a nurse, a dietician or a pharmacist. Each patient would be assigned an advocate who would know the patient's case and carry a beeper 24 hours a day to answer questions. These are just a few of the differences between existing care in the market and what this focused clinic would do.
From a financial perspective, the design was radically different as well. The doctors were not getting paid on a fee-for-service basis, but instead were paid on a salary with no incentive to over utilize. In order to determine how this model was performing vs. traditional care, an administrative process was set up to compare the two. This was designed so the biller would create a dummy bill for each A-ICU patient. The patient would have no financial responsibility for the bill, but it would go to the TPA so that analysis could be done regarding the total cost of care. It was felt that not all the targeted patients would sign up to be seen in the clinic and the ones that didn't would act as the control group from a cost standpoint. The team would compare the total cost of care in both groups and if the new process was successful it should produce a lower total spend.
One of the problems with this model was that the patients using the clinic for heart disease or diabetes care might also have other complaints that needed to be worked on. The solution here was to send a live bill for a diagnosis that was other than heart disease or diabetes. For example, if the patient injured a knee that would be treated outside of the process for which they had signed up. The clinic wouldn't turn them away, but the employer would have to pay for anything outside of the targeted diseases.
A business model for the clinic was developed which included the afore mentioned staff and simple clinical space. The initial intent of the business coalition was to fund the clinic as a stand-alone expense. This was designed assuming there would be significant savings with the new model of care and the employer would benefit depending on the magnitude of these savings. Once the start-up expenses were covered for the clinic, a 50/50 split of the savings would be shared between ThedaCare and the coalition. A proforma was developed for the clinic, which estimated the cost would be about $900,000 for the first year of operations. This would include a full-time physician, nurse and diabetic educators, part time pharmacist, social worker and full-time administrative staff to manage the people and accounting processes.
The key to the success of the model was the number of patients seen. The proforma (fig.2.) showed the employers would save nothing if 500 or fewer patients were seen. The savings increased significantly as when more than 500 employees participated. If a 1,000 patients were seen, a million dollars of savings could be achieved, which increased exponentially after that. Most of the clinic's cost is a fixed expense that is required to deliver the new clinical process. This is why it was imperative to have a large enough volume of patients to drive enough savings to offset the expense.
Year One Reimbursement Model Example
|Minimum Participation of
500 Pts Achieved
|Minimum Participation of
500 Pts NOT Achieved
|Annual Gross Savings ($4,000/member/year)||$4,000,000||$1,800,000|
|Annual Operating Expenses**||$1,100,000||$1,100,000|
|Savings Less Expenses||$2,900,000||$700,000|
|Minimum Savings to BHCG||$300,000||$0|
|Net Savings (Gross Savings - Ops Exp - BHCG Min)||$2,600,000||$700,000|
|Net Savings Split 50.50
Share Payout to BHCG
Share Payout to ThedaCare
** Year One operating expenses include a one-time equipment expense of $150,000
Fig. 2. Financial proforma based on different scenarios.
The concept for reimbursement was to set an annual savings target. This was established as actual operating expense of the clinic plus $300,000.The “gross savings” = control group expenses — participation group expense for the same period.
The “net savings” = gross savings — (actual operating expenses - $300,000). If the minimum participation level of 500 patients were met net savings would be equally divided on a 50/50 basis between the coalition and ThedaCare.
If the minimum participation was not met, ThedaCare and the coalition would split the gross savings — operating expenses on a 50/50 basis. The purpose of the $300,000 was to make sure some savings would accrue to the employers for their $950,000 investment and to give ThedaCare a minimum target to manage to before they received any payment.
There are many other details to the financial model that, if interested, the ThedaCare Center for Healthcare Value can share with interested parties.
In addition to the financial targets, there was also agreement that the quality targets described above would have to be met before any pay out occurred. In addition, it was agreed an independent auditor would be engaged to monitor performance and a steering committee of the stakeholders who had participated in the value stream would be formed to address issues that were sure to come up over a three-year period.
So what happened?
The concept was taken to the chief executive officers of the business coalition who had several concerns. The most important one was they didn't know how to educate employees to use the clinic. If they couldn't get their employees to go, there was no chance for savings. In fact as this concept becomes more popular, other markets are finding similar issues. If the patients won't come, there is no way to drive improvement. There will need to be significant steerage built in to the benefit plans to make this happen. Simply removing co-pays probably won't accomplish this. In chronic disease patients, they may need more positive incentives or removal of benefits if they don't comply. We tend to like positive incentives such as free drugs, athletic cub memberships, etc. as ways to entice people to participate.
Another concern was that the A-ICU might conflict with programs some businesses already had in place. At least two employers had already contracted with onsite clinic providers and were beginning to address their employees' health needs directly. There was fear that the A-ICU could jeopardize the effort they were already funding.
There was discussion among the coalition members that the A-ICU would set off a backlash in the existing provider community that they didn't want and had little energy to deal with. There was concern ThedaCare would be seen as the enemy by providers in the market and costs might actually increase because existing providers, who had a monopoly on hospital and surgery services, would increase rates collectively.
Finally, this proposal was presented at the end of October 2008 during the worst of the financial meltdown and leaders were concerned about adding any cost from a healthcare standpoint even if it had a potentially huge upside.
Where this effort stands at this point is unknown. There still is an interest among some of the coalition's members to do a clinic. It may not be a full-blown A-ICU model if there is not enough interest to generate at least 500 patients. The learning at this point is that without committed employers willing to take risk and promote radical change in a market, we can continue to expect more of the same — rising cost and no improvement in quality.