Task Force Report
Incentives for Enhancing Stroke Care
|Sandra M. Schneider, M.D.
Task Force Chair
Strong Memorial Hospital
University of Rochester Medical Center
|Anthony J. Furlan, M.D.
Cleveland Clinic Foundation
|Larry B. Goldstein, M.D.
Task Force Co-Chair
Duke University Medical Center
|Michael D. Hill, M.D.
Foothills Medical Center
|James G. Adams, M.D., F.A.C.E.P.
|Michael T. Rapp, M.D., J.D.
George Washington University
|Kenneth L. DeHart, M.D., F.A.C.E.P.
Carolina Health Specialists
Myrtle Beach, South Carolina
|Joseph P. Wood, M.D., J.D.
Mayo Clinic Hospital
|Andrew M. Demchuk, M.D.
Foothills Medical Center
Health care providers strive to provide optimal treatment for their patients. However, medicine is becoming more complex and is practiced in a multifaceted environment that includes an array of potential barriers to effective acute stroke care. The identification of appropriate incentives to overcome these impediments is critical to advancing the overall level of stroke care in the country.
The emergency department (ED) plays an important role in achieving optimal care of the stroke patient. It provides immediate accessibility, integration with the emergency medical services (EMS) system, and access to a hospital's resources. A stroke patient's initial contact with hospital-based personnel is generally in the ED.
Until the advent of treatment with intravenous t-PA, ischemic stroke was considered not reversible even by aggressive emergency care. The goals of acute treatment were to avoid complications and reduce the risk of recurrence. Some of the barriers to optimizing acute stroke care are related to general challenges faced in this care setting. The ability to deliver optimal care can be affected by limited available resources. For example, a sudden increase in patient volume can make it difficult to meet the immediate needs of all patients. Because stroke patients often present with limited readily available background history (e.g., pre-existing conditions, current medications, etc), hospital personnel must take the time to gather this important information. A single patient requiring aggressive treatment and the attention of multiple staff members can exacerbate this problem.
Compounding these problems, nearly 70 percent of ED care is delivered in "off-hours" (ie, evenings, nights, weekends, and holidays), times when consultative support is even more limited. Therefore, the emergency physician must often evaluate patients with neurological complaints without the benefit of consultation with a neurological specialist. This can be a particular barrier to the delivery of therapies with a narrow therapeutic index such as intravenous thrombolysis. Similarly, radiology resources in many hospitals can be overloaded during the off-hours when technicians and radiologists are often in short supply. Smaller EDs may not have immediate imaging resources available at any time. These types of challenges have increased over the past few years because of ED overcrowding. More than 90 percent of ED directors perceive that their departments are either at or over capacity (1,2), and available data indicate that this perception is well founded. According to a recent American Hospital Association poll, 62 percent of hospital EDs are at or over capacity. An ED point prevalence study done on a typical spring evening found there were 1.1 patients per treatment space, 4.2 patients per registered nurse, and 9.7 patients per physician (3). This overcrowding introduces stress in the system that can interfere with the optimal delivery of care.
Given the nature of emergency care, emergency physicians must frequently rely on consultation with physicians from other specialties. These physicians assist in evaluation and treatment and assume the care of patients requiring hospital admission. The importance of consultative services to a smooth functioning ED is recognized in the EMTALA (Emergency Medical Treatment and Active Labor Act) statute that requires hospitals to identify specific physicians with on-call responsibilities. This has become more difficult in recent years (see below, Neurology section).
Using intravenous t-PA as an example of the problems associated with the use of an innovative therapy with a potentially narrow therapeutic index, most emergency physicians have been unwilling to accept sole responsibility for a decision to administer the drug without adequate consultative and administrative support. Concerns focus on the need for expert CT scan interpretation and detailed neurological examinations with continuing patient reassessments. These concerns are heightened in the increasingly hectic and crowded ED environment. In addition, relative lack of extensive experience with the use of t-PA leads most emergency physicians to seek the opinion of a neurologist prior to drug administration, or even to require that the neurologist order the treatment. In contrast, ED physicians are quite willing to independently administer t-PA for myocardial infarction. Unlike CT scan interpretation, identifying ST elevation on an EKG is one of an emergency physician's basic skills. The indications for t-PA for myocardial infarction are more straightforward, the frequency of its use in this situation is greater, the risk to the patient with treatment is less, and there is less controversy regarding its use.
In the United States, the problems of lack of space, personnel, technological resources, and support from subspecialty consultants are formidable, and the situation is much the same in Canada. These barriers can be partially addressed through development of efficient prehospital and ED triage systems, increased funding to support the availability of needed technological resources in hospitals caring for patients with acute stroke, and the development and adoption of care pathways including predetermined orders to facilitate the use of potentially hazardous stroke therapies by emergency physicians. However, these approaches do not address the need for consultative support.
As reflected above, a perceived or actual lack of consultative support is viewed by emergency physicians as a barrier to the provision of potentially risky therapies, such as thrombolytics for acute ischemic stroke. As compared to acute cardiac emergencies, many emergency medicine physicians are not sufficiently trained in neurology to feel comfortable using new stroke therapies that have the potential to cause harm.
Although stroke is one of the most common inpatient problems confronting neurologists, many neurologists subspecialize and do not provide care for stroke patients on a regular basis. Beginning in 1996, most neurology residency programs in the United States began including some training related to intravenous t-PA for acute stroke. Nonetheless, many neurology residents have little experience or confidence in the use of intravenous t-PA (4). Surveys in Texas (5) and by the American Academy of Neurology (6) suggest that no more than 50 percent of American neurologists have given intravenous t-PA for acute stroke. Less than one third (30 percent) of the neurologists found the evidence for t-PA efficacy "very convincing," with the majority (67 percent) finding the evidence of efficacy "somewhat convincing." Many neurologists felt the drug was "too risky," and 62 percent were "very concerned" about intracerebral hemorrhage.
In addition to uncertainty about the efficacy of new therapies, most general neurologists see patients in outpatient settings. The economics of clinical practice dictate a tightly scheduled day. Interruptions to evaluate a patient with acute stroke can take several hours. Given the limited financial reimbursement associated with this activity (see below, Financial Reimbursement section), there is a strong disincentive to leaving a crowded office to provide emergency consultative services. The acute stroke time targets developed at the first NINDS National Symposium in 1996 included "access to stroke expertise" within 15 minutes of ED arrival (7). The phrase "access to stroke expertise" was purposefully used because neurologists are not immediately physically available in many community or rural hospitals. Therefore, the recommendation anticipated telephone consultations, telemedicine, and the fact that in many situations it may be impossible to have a neurologist physically present in the ED before emergency therapies such as t-PA are started. Although consultants are legally liable for advice given over the telephone, there is no financial reimbursement for telephone consultation. Therefore, many neurologists refuse to give advice about intravenous t-PA over the telephone. Finally, when radiologists are not immediately available, neurologists interpret radiographic studies such as CT scans to guide treatment. However, they are rarely financially reimbursed for these activities. Given this reality, it is not surprising that emergency medicine physicians cite a lack of neurological support as a barrier to acute stroke treatment.
The introduction of t-PA for acute stroke led to several legal theories that could provide a basis for a claim for medical malpractice, defined as violation of the accepted standard of care causing harm to a patient. Similar theories could apply to other innovative acute stroke therapies. The primary basis for any malpractice claim is a bad outcome. With a bad outcome, the task from the plaintiff's perspective merely becomes relating the outcome to a deviation from the standard of care. Evidence of the standard of care is provided by the opinion of expert witnesses, with arbitration by a judge and jury. One physician with minimal qualifications expressing an opinion that certain conduct violated the standard of care can be sufficient to have a jury consider the case and award damages to compensate for the bad outcome.
In the case of any innovative stroke treatment, the starting point is often a bad outcome (because the patient had a stroke). The first theory available to a plaintiff's attorney would be to argue that failure to administer t-PA was a violation of the standard of care, and that, had it been used, the outcome would have been the elimination of the patient's neurological deficits. The latter is hard to prove scientifically, but easy to establish in a court of law since it merely requires the opinion of a qualified witness. A second theory available to a plaintiff would be to argue that the administration of t-PA was either not indicated or that the drug was improperly administered, leading to a bad result (either hemorrhage or perhaps simply failure to be cured). Finally, there is the doctrine of informed consent. In this case, a plaintiff's attorney could argue that the patient should have been informed of t-PA as an available treatment including its risks and benefits.
The Food and Drug Administration approved intravenous t-PA as a treatment for acute ischemic stroke in 1996. Shortly thereafter, its use was advocated in a scientific statement from the American Heart Association and an almost identical practice parameter from the Quality Standards Subcommittee of the American Academy of Neurology (8,9). Editorials were published and educational campaigns were launched, aimed toward educating physicians and patients alike. The popular press, magazines, newspapers, and media of all types heralded the new treatment for stroke, sometimes overstating the therapeutic potential.
These efforts in part created a fertile field for malpractice litigation related to the use of intravenous t-PA. The plaintiffs' bar recognized that a new potential cause of action existed. Advertisements and websites for malpractice attorneys highlighted the "alarmingly low" use of t-PA for patients with acute stroke, "especially for African Americans" (10). Solicitations educated patients, "If you suspect that a loved one should have received t-PA but did not, or that t-PA was administered improperly, it may be important to contact an attorney" (11).
Legal solicitation continued as medical scientific debate and physician education was ongoing. Legal scrutiny was noted even before there was national consensus among specialties. With such a threat present, establishing national consensus to undertake a time-sensitive, technically demanding, high-risk therapy was unlikely. The scientific evidence was argued to be insufficiently convincing, the operational challenges enormous, and the complication rate worrisome. The perception was that any effort to further advocate for the use of thrombolytics could increase legal risk. The counter argument, whether true or not, was that a higher threat existed from failure to administer thrombolytics. This argument was not persuasive, but was frightening and believable.
There are several ways to deal with high-risk medicolegal situations, such as those involving t-PA. From the standpoint of emergency physicians, appropriate consultative support for treatments with a narrow therapeutic index is viewed as important. The development of institutional policies for the use of a specific treatment also reduces medicolegal risk. For example, an institutional policy decision could define the circumstances in which a treatment could or could not be safely administered. Such advance policy decisions, if reasonable and if followed, can provide protection against the opinion of a plaintiff's expert. Support may also be obtained from guidelines or policy statements by professional organizations. When a real difference of opinion exists within the medical community, an expert's adoption of one view does not lead to legal liability.
Emergency physicians indicated to the American College of Emergency Physicians that they were facing litigation for failure to administer t-PA for acute stroke. Given the perceived medicolegal risk, the American College of Emergency Physicians developed a policy on the matter (12). This policy statement indicated (in part) that, "Intravenous t-PA may be an efficacious therapy for the management of acute ischemic stroke if properly used incorporating the guidelines established by the National Institute of Neurological Disorders and Stroke (NINDS). There is insufficient evidence at this time to endorse the routine use of t-PA in clinical practice when systems are not in place to ensure that the inclusion/exclusion criteria established by the NINDS guidelines for t-PA use in acute stroke are followed. Therefore the decision for an ED to use intravenous t-PA for acute stroke should begin at the institutional level with commitments from hospital administration, the ED, neurology, neurosurgery, radiology, and laboratory services to ensure that the systems necessary for the safe use of fibrinolytic agents are in place." Similarly, the Canadian Association of Emergency Physicians published a policy indicating that current evidence did not support the use of t-PA in acute stroke as a "standard of care." The American Academy of Emergency Medicine was the first professional association in the United States to note that "evidence...is insufficient to warrant [t-PA's] classification as standard of care." This announcement served as powerful advocacy and welcome support for emergency physicians who felt great medicolegal threat, even while trying to provide the highest levels of service to patients (13).
The experience with t-PA may be viewed as a model of what can be expected for other emerging acute stroke therapies in the future. Proponents of the use of innovative therapies for acute stroke must frame a convincing scientific stance and ensure the high levels of system supports that will reliably achieve timeliness and safety. Promulgation of new, potentially hazardous acute stroke therapies before health care providers and institutions have been educated about appropriate organizational changes and support mechanisms will likely bring about challenges similar to those facing t-PA use.
Appropriate financial support for consultative services and stroke systems has been discussed as important to the provision of optimal stroke care. In addition, indigent care must be supported. Because there may be different payers for acute and long-term care, even if an acute treatment is cost-effective from a societal standpoint, it may increase the costs to those providing the treatment. Other health care systems avoid this conflict through global health care budgeting. The financial disincentives to the use of intravenous t-PA would be similar for other innovative approaches. For example, hospital and physician reimbursement is the same whether or not t-PA is administered to an acute stroke patient. Unless payers recognize the added value of implementing an acute stroke response system, many hospitals will be unable to dedicate the resources needed to consistently provide rapid, state-of-the-art care to acute stroke patients.
Surprisingly, only a minority of neurologists listed low reimbursement as a reason for not giving t-PA. This contradicts the common knowledge that neurologists consider reimbursement for intravenous t-PA inadequate, especially when they must go to the hospital in the middle of the night, on weekends, or during a busy daytime practice.
Many neurologists are unfamiliar with optimal current procedural terminology (CPT) billing for thrombolytic therapy. The American Academy of Neurology has published a Stroke Coding Guide, which advocates using CPT Codes 99223 (initial hospital care, high complexity), 99291 (critical care, first hour), and 99292 (critical care, subsequent half hour) for initial treatment of acute stroke patients with thrombolytic therapy. The average national payment in 1997 using these three CPT codes was $394.77; in 2002, the average payment increased to $447.43. Although there is a CPT code for intravenous t-PA for acute stroke (37195), the work relative value unit (RVU) for this code is 0! This reimbursement rate provides little incentive for neurologists to interrupt an otherwise busy workday or to respond during evenings, weekends, or holidays. Recognizing that it is unlikely that the emergency physician and neurologist will both be fully reimbursed when they bill under the same codes for the same patient, there is little financial incentive to establish a highly functional team approach.
Adequate financial resources to promote optimal stroke care and to support physician leadership for the direction of a multidisciplinary stroke team is critical. A significant opportunity exists to amend the current system to create meaningful incentives in the provision of contemporary stroke care.
From the institutional standpoint, governmental payers such as Medicare compensate facilities based on a diagnosis related grouping (DRG) methodology. This DRG methodology largely reflects overhead costs calculated from "case data" with little recognition of the expense and higher resources involved in contemporary thrombolytic or evolving neurological therapies and technologies. Commercial payers typically compensate acute care and rehabilitation facilities on a per diem basis, with denied payment inconsistency occurring for services involving progressive stroke care. These reimbursement limits create a disincentive to the provision of optimal stroke care. New compensation levels must reflect three essential elements: 1) the increased costs involving present and evolving thrombolytic and novel interventions; 2) the cost in providing uncompensated care; and 3) resources to develop and support medical leadership and system analysis.
The support of a "stroke team" and team leader is particularly important. Traditionally, medical directorships have been developed for care that typically involves specific units within a given hospital. Stroke care, however, occurs in many locations within the hospital. It will be difficult, if not impossible, to nurture an optimal approach to care without financial resources to properly encourage physician leadership and system development with meaningful clinical quality review.
Incentives must also be developed for those providing primary bedside care. One opportunity involves amendments to the current nomenclature involving the CPT-4 codes used by physicians and ancillary providers to describe the care provided to stroke patients. A second opportunity involves the development of new CPT codes to accurately reflect new services in progressive stroke care.
The CPT-4 published by the American Medical Association is the prevailing nomenclature in the United States used to describe provider services to patients and it includes evaluation and management visit services, observation services, and critical care services. However, "concurrency of care" limitations in the provision of services occur when they are simultaneously provided by multiple physicians from multiple specialties. These types of restrictions undermine an integrated multidisciplinary approach and must be discontinued.
CPT codes presently exist that could be helpful in promoting an integrated team approach. They include: 1) 99360, physician standby service, requiring prolonged physician attendance, each 30 minutes; and 2) 99371, telephone call by a physician to a patient, for consultation or medical management, or for coordinating medical management with other health care professionals (e.g., nurses, therapists, social workers, nutritionists, physicians, pharmacists). Other existing codes include: 1) 99371, simple-brief, e.g., to report on tests and/or laboratory results, to clarify or alter previous instructions, to integrate new information from other health professionals into the medical treatment plan, or to adjust therapy; 2) 99372, intermediate, eg, to provide advice to an established patient on a new problem, to initiate therapy that can be handled by telephone, to discuss test results in detail, to coordinate medical management of a new problem in an established patient, or to initiate a new plan of care; and 3) 99373, complex or lengthy, eg, lengthy counseling session with anxious or distraught patient, detailed or prolonged discussion with family members regarding seriously ill patient, or lengthy communication necessary to coordinate complex services of several different health professionals working on different aspects of the total patient care plan.
The majority of third-party payers in the United States, including Medicare, do not typically recognize these physician services for payment. Efforts should be directed to securing specific descriptive language amendments in these codes in order to support the development of multidisciplinary teams and to recognize the challenges in providing consultative services to acute care providers and patients.
The second opportunity for developing incentives exists through the creation of new and innovative language within CPT-4 to specifically describe acute resuscitative and chronic stroke care. This descriptive nomenclature for acute care could be global in scope and analogous to the present Critical Care 99291 or the CPR 92950 codes. As an example, an entirely new CPT code could be developed that more accurately and fairly captures the physician resources in the provision of stroke care.
Some of these financial incentives have already been implemented in Canada over the last 3-5 years. For example, there has been a general move toward payment for on-call services, with several provinces providing stipends for either neurology call (including acute stroke) or acute stroke only. Payments range from $150 a day (neurology including acute stroke care) in Manitoba, to $300 a day (stroke call) in Alberta, to $500 a day (stroke only) in Ontario, to as high as $740-$850 a day (neurology including acute stroke care) in British Columbia. In addition to billing for their usual consultation fee, most bill for an additional 30-90 minutes for the time required for t-PA administration.
A coordinated stroke reimbursement strategy by patient advocates and professional organizations may be the most effective way to change the reimbursement system to facilitate quality stroke care. Nationally, the Centers for Medicare and Medicaid Services (CMS) regularly reviews reimbursement levels for specific conditions and physician payments and invites public comment. Locally, carrier review committees can be similarly approached.
Systems approaches to stroke care are central to many of the identified resource issues. Studies show that having an organized system of stroke care improves outcomes, reduces complications, shortens hospitalizations, and reduces costs. However, current reimbursement systems in the United States do not support this approach. In addition, reimbursement levels have not accounted for the added expense involved in the use of new therapies. For example, DRG14 (stroke) was created before intravenous thrombolysis was introduced.
Stroke quality assessment was recently discontinued by the CMS. As a result, there is currently little incentive for hospital systems to monitor the quality of their stroke care. Programs to identify stroke centers are being discussed, and CMS is planning to reintroduce stroke quality indicators in its next scope of work. This should help in gaining institutional support to enhance stroke process improvement programs and provide a consistent mechanism for giving feedback to all caregivers, including emergency care providers.
Incentives to Improve Acute Stroke Care
Return to Table of Contents
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Bethesda, MD 20892
Last updated April 28, 2011