Facing escalating costs of medications and technology, health care patients and providers in the United States continue to search for opportunities to reduce overall costs while maintaining and improving health care outcomes. At the Mayo Clinic Comprehensive Stroke Center Practice, we conducted a project to design and deliver care more customized to the needs of individual patients while reducing cost and resource constraints. It is a risk-stratified approach that could be applied to treating many medical conditions.
The Mayo Stroke Practice used time-driven activity-based costing (TDABC) to study costs associated with alternative protocols for stroke care (see the graphic below). TDABC uses a bottoms-up approach to identify the actual clinical processes and resources used to care for a patient over a period of time. TDABC works from a process map of a patient’s care pathway, attributing costs to the time of each resource used at each step of the pathway. With this information, clinicians learn how to make more efficient use of high-cost resources, leading to lower total costs while achieving the same or better patient outcomes.
However if one could predict a patient does not need such care, this could save the system, as well as payers, a lot of money. The daily cost of an NSPCU bed — both to the payer (insurers and patients) and the hospital — averages $500 a day less than an ICU-level care bed, which is multiplied by length of stay (LOS). There are also measurable costs “turning over” a hospital bed in terms of both time delays such as patient’s waiting in the emergency department as well as financial expenses in cleaning and sanitizing a room to be ready for the next patient. Similar to the opening moves in a game of chess, which can determine the rest of the game, similar bed logistics can make or break hospital bed flow. So how can one improve both hospital bed flow and improve value-based care in stroke patients?
Using existing stroke data and TDABC mapping, one can stratify a stroke patient’s true risk for needing or not needing ICU-level care using the National Institutes of Health stroke scale (NIHSS). Historically, a “step-down” unit or progressive care unit (PCU) was typically used on the back end after ICU-level care for patients too sick and unsafe to send to a regular hospital bed because they might decompensate and end up back in the ICU. Using a NPCU strategy on the front end for some stroke patients is revolutionary in the sense patients are admitted directly from the emergency department after receiving TPA. This reengineering of hospital bed flow allows a relative cost savings without compromising quality and improves the value.
Since 1995, when TPA was FDA-approved to treat stroke patients, the common practice was to monitor these patients in the ICU environment due to concerns for decompensation from intracranial bleeding and complex interventions. Under fee-for-service reimbursement, however, stays in the ICU can incur daily charges up to $2,500, nearly 25% of Medicare’s total reimbursement ($11,000) for TPA treatment.
The Mayo stroke team used the NIH Stroke Scale (NIHSS), which ranges from 0 (normal) to 42 (severe), to stratify patients into different risk categories and identify those who truly needed ICU-level care. In a trial for intravenous TPA for acute stroke care, reported in 1995 in the New England Journal of Medicine, the average NIHSS score was about 14.
The most severely affected stroke patients had a NIHSS greater than 24 were most likely to need ICU-level care for monitoring. Therefore, Mayo Clinic’s stroke center data showed similar findings and proposed that stroke patients with an NIHSS score of 18 or higher should be monitored in the ICU for the first 24 hours after receiving TPA. Such patients often suffered medical complications that required advanced interventions such as intubation and mechanical ventilation. However, patients with few comorbidities and NIHSS scores of 14 or less had a reduced probability of severe complications that required critical interventions. Care for these patients could potentially be managed and monitored in the lower-cost NSPCU environment.
The team saw an opportunity to reduce costs based upon how and where patients received care, while still meeting Joint Commission requirements for post-TPA care, by treating low-risk patients in a NSPCU-level bed with a specialized hybrid level of nursing care (see the table below) for the first 12 hours. This risk-stratified care model improved value by delivering equivalent care quality with a lower-cost mix of resources. In addition, the stratification process allowed for better “demand elasticity” of ICU bed utilization.
Comparing NeuroICU and NSPCU Nurse Monitoring for Stroke Patients
Optimizing NSPCU and ICU bed utilization therefore is analogous to the game of Tetris in which players fit blocks of various sizes inside an available structure. All hospitals play a similar game to optimize space utilization by getting the “right patient to the right bed” with the fewest moves possible. ICU-level care beds are the most expensive in the hospital and are reimbursed at the highest rate. Ideally, they should be used only for the most complex medical/surgical cases or for transfers from emergency department (ED) and other hospitals. By freeing up ICU beds, previously used for lower-risk stroke patients, hospitals have more capacity, or elasticity, to admit postoperative ICU patients and ICU admissions from the ED and allow those care teams to focus on those patients.
Getting the right patient to the right bed also reduces the number of transitions of care (TOCs). Historically, some patients underwent four separate handoffs as they made transitions initially from ED or an operating room to the neuroscience ICU, then to the NSPCU, and finally, to a regular floor bed. This represents at least four moves (A →B→C→D) for the patient and adds risks: Details about medication allergies and other Important information about the patient can be lost, communicated incorrectly, or misconstrued during the handoff from one care team to the next. Handoffs are similar to those in football. The number of handoffs increases the complexity of the play and is associated with a higher likelihood of “fumbles,” or medical errors. When stroke patients are admitted from the ED directly to the NSPCU, a regular floor bed the next day, then discharged home, there is at least one less TOC, or handoff. In addition to reducing the total number of TOCs, a standardized, or structured, communication tool — a checklist — for exchanging important patient information during handoffs can reduce the number of medical errors as well.
As illustrated in the above examples, the ability to stratify and predict patient needs up-front opened the door for actions that enhanced process efficiencies, reduced operational costs, and improved patient outcomes. Patients that received TPA and were subsequently monitored in the NSPCU had an average reduced cost of 25%. Of 448 stroke patients seen in the past three years, all of whom would previously been sent to the ICU, 166 (37%) were monitored in the NSPCU, leading to a net cost reduction of nearly 10%, with no adverse impact on patient outcomes.
While the role for a progressive care, or step-down, unit is not new in health care, it is one we believe may be underutilized for elderly and more complex patients, especially when its cost advantage over the highly-resourced ICU has not been quantified. An NSPCU increases the effective capacity of existing ICU-level beds and provides better utilization of regular-ward-floor beds for medically-stable patients. Importantly, the risk-stratified approach does not replace or supersede physician judgment about factors not accounted for in the NIHSS-weighted model when deciding the best overall course and bed status for the patient.
As this case illustrates, process mapping of care pathways and accurate costing makes it possible to design and deliver care that is more customized to the needs of individual patients. The customization produces equivalent or better quality and outcomes at reduced costs because of more efficient resource utilization and diminished risk from medical errors. None of the gains discussed in this article are unique to stroke treatment, and the NSPCU model can be extended to many medicine and surgery areas to improve the value delivered at hospital, national, and international levels.