Overview of the Study
This multi-center, randomized, controlled trial is designed to evaluate a standardized, educational intervention in increasing appropriate tPA use in stroke. The study is based on adult education and behavior change theory and consists of three phases.
Purpose and Background Information:
The purpose of the study is to determine if a standard approach to assessing tPA treatment barriers, combined with a series of educational interventions directed towards hospital systems, focusing on the Emergency Department and Emergency physicians, will be successful in: 1) increasing the number of acute stroke patients treated appropriately with tPA in the Emergency Department and 2) improving ED physician knowledge and attitudes toward tPA use in stroke. Stroke is the #1 cause of disability and the #3 cause of death in the US. There is only one FDA approved treatment for acute ischemic stroke, tPA (tissue plasminogen activator, a “clot busting” drug) that must be given within the first 3 hours of stroke onset. Only 1 to 3 percent of stroke patients receive tPA treatment in emergency departments due to a number of patient and system related issues. Previous studies have shown that improved education to physicians can increase appropriate tPA use in the emergency department and potentially improve stroke patient care and recovery. An extensive literature review is available in the protocol.
Inclusion/Exclusion Criteria:
Twenty four hospitals were identified from a pool of all Michigan acute care hospitals located in the lower peninsula. Selected hospitals met the following inclusion criteria: physician-staffed Emergency Department at all times, 24/7 CT scanning available, computerized pharmacy dispensing system or thrombolytic use log, agreement to participate and identified site investigator. Within each of the 24 hospitals, all ED physicians were invited to participate. Exclusion criteria included: hospitals that care primarily for children, psychiatric patients, and long-term care (convalescent) facilities, hospitals with well-established comprehensive stroke care centers, Emergency Departments with an annual volume of greater than 100,000 patients (only one hospital met this criteria), hospitals with less than 100 stroke patients treated each year, hospitals in the Upper Peninsula of Michigan, and physicians outside of the ED were excluded due to infrequent participation in the acute treatment of stroke patients.
Study Population:
1. ED Physicians: Approximately 400 total ED attending physicians are participating at 24 sites throughout the lower peninsula of Michigan. Pregnant female emergency room physicians may participate in this study as it is of minimal risk to them or to their fetus.
2. Retrospective chart/imaging reviews: There is no specific active recruitment of patients for the retrospective review. Patient records identified for retrospective review are identified through monthly pharmacy tPA use reports. The site investigator checks the chart of each tPA treated patient to verify treatment with tPA for acute ischemic stroke in the ED.
Intervention:
The study consists of 3 phases:
A pre-intervention phase establishing:
The baseline proportion of patients with ischemic cerebrovascular disease treated with thrombolytic therapy using ICD-9 data and pharmacy records of Emergency Department tPA usage. Twelve months of baseline data was obtained.
Emergency physician knowledge and attitudes regarding thrombolytic use in stroke. Knowledge will be assessed via a web based survey. All ED physicians at each of the performance sites (approximately 400) were invited to participate.
Random assignment of performance sites to either “intervention” or “control” arms.
An intervention phase (“intervention sites only) consisting of:
Identifying and supporting “local stroke champions”;
Conducting a 1-day study start-up and educational meeting (“Champions meeting”) in Ann Arbor
Conducting on-site assessments of barriers to tPA use in stroke. Barriers will be identified using focus groups of ED physicians and other ED staff, interviews with hospital administrators, neurologists and radiologists and a walkthrough of each ED focusing on site resources.
Providing two 1-day educational interventions (at 3-month intervals) consisting of on-site CME sessions for medical staff and nursing personnel (2 hours) and a mock stroke “code” tailored for specific dentified barriers.
Additionally, the following mechanisms will be provided to enhance behavior change in the intervention group: providing physicians access to thrombolytic stroke expertise as needed (24/7) via the University of Michigan INSTINCT Group on-call system; providing electronic messaging feedback on hospital thrombolytic use; and providing critical incident debriefing for known tPA-use complications (ICH)
The post-intervention phase, (all sites) will consist of:
Monthly assessment of tPA use through pharmacy use reports
Quarterly review of ICD-9 reports,
Review of all ED tPA treated cases by retrospective chart/imaging review. The focus of the reviews is appropriateness of tPA use, compliance with established tPA treatment guidelines and tPA-related complications. Preliminary reviews will be conducted by site PIs (at intervention sites only) and results forwarded to the clinical coordinating center along with a de-identified copy of the medical record and neuroimages. A second, blinded record review is done by 2 separate physician investigators. A comprehensive chart review is completed by a trained nurse study coordinator.
Re-assessment of physician attitudes and knowledge regarding thrombolytic use in stroke at 12 months following the second interactive educational intervention.
Outcome measures include:
Changes in tPA use rate, with assessment of appropriateness of use and complications.
Changes in emergency physician knowledge and attitudes regarding thrombolytic use.
Data Collection:
To assess physician knowledge and attitudes regarding tPA, the ED physician survey is administered using a Web-based tool (Survey Monkey) at 2 time points during the study. Surveys take approximately 15 minutes to complete. Paper copies of the survey are provided for any physician who indicates such a preference. Physicians who do not complete the survey receive 2 email reminders and a follow up contact from the site PI or coordinating center staff.
To determine significant facility change over the course of the study, the local site PI completes a one page facility survey (Annual Hospital Characteristics Survey) yearly. Information includes: number of hospital discharges, current number of inpatient beds, ED volume, JCAHO stroke center status, site CT, neurology and neurosurgery support availability and advanced stroke treatment (Intra-arterial tPA) capability.
To determine the proportion of acute ischemic stroke patients treated with tPA, site pharmacy departments produce reports of tPA use. Computerized pharmacy dispensing systems are capable of producing these reports and are used if available. Billing records are used at sites without computerized dispensing systems. Data was obtained for the 12 months prior to start of the study (to establish baseline treatment rates) and then monthly during the study. Medical records departments provide reports of the number of stroke patients treated at the facility using pre-identified ICD-9 codes. Data was obtained for the 12 months prior to the start of the study (to establish overall stroke admission rates) and then quarterly during the study. Reports contain only the number of patients having been assigned stroke related ICD-9 codes. No patient identifiers are obtained.
To determine the appropriateness of tPA use, a retrospective chart/imaging review is conducted on records of all stroke patients treated with tPA during the study. Patients are identified using the pharmacy tPA use reports (#3 above). The local site PI verifies that tPA was used for acute ischemic stroke. At intervention sites, the local site PI confirms tPA eligibility criteria and assesses for important clinical events, using a brief (2 page) chart review tool. The tool contains only a study specific patient ID code. The completed “Site Principal Investigator Chart Review” tool is forwarded along with a de-identified copy of the medical record, to the Coordinating Center. Control sites will provide only a de-identified copy of the medical record. A review of the de-identified medical record will be conducted by 2 blinded physicians who are stroke treatment experts not directly involved in the study. Nurse coordinators from the Clinical Coordinating Center will conduct a final, more extensive chart review, obtaining additional detailed information on demographics, medical history, details of the treatment process and patient outcome at discharge (clinical status and Modified Rankin Scale). This information will be collected on the “Nurse Coordinator Chart Review” form.
Safety of tPA use is an important secondary outcome measure in the INSTINCT protocol. Due to variability in the description of written radiology reports regarding the presence or absence of post-treatment intracerebral hemorrhage (ICH), the actual images (de-identified by removing the patient’s name and hospital ID#) of all brain CTs and MRIs obtained during the first 10 days following treatment are obtained for blinded review. Images are copied to disc and each disc labeled only with the patient’s study ID#.
Determining post treatment clinical events including outcome and treatment complications of patients treated with tPA is critical to the study. It is imperative that a complete medical record be available to the INSTINCT team. It is possible that tPA treated patients may be transferred away from INSTINCT study sites during the acute treatment of the index stroke for a variety of reasons, for example no available ICU beds, lack of neurosurgery support, need for specialized care not available at the site, etc. In these cases, the medical record from the INSTINCT site is incomplete. The INSTINCT study will make all reasonable efforts to obtain the medical record and CT/MRI images from the receiving hospital.
All applicable HIPAA regulations are honored. The INSTINCT study is being conducted under a waiver of written informed consent and a waiver of privacy authorization for the retrospective reviews. Each participating institution has also executed a Data Use Agreement allowing the release of data meeting the definition of a limited data set. Specific requirements of receiving hospitals will be determined and met.
For hospital sites randomized to the “intervention” arm, an on-site assessment will be completed with each ED to identify treatment barriers. Information about barriers will be generated from structured interviews with key site personnel, a focus group with ED physicians, a focus group with ED nursing personnel and other key clinical personnel, and a site tour evaluating systems and resources for acute stroke treatment.
Data Analysis:
The primary comparison will be between the change in proportion (percentage) of acute ischemic stroke patients treated with tPA during the four years of the study. Change in physician knowledge and attitude toward tPA use over time will also be analyzed using the electronic survey results. Data analysis is being conducted by the University of Michigan School of Public Health Epidemiology and Biostatistics Data Management Core.
Risks of the Study:
Risks of participation in the study are considered to be “no more than minimal”. Risks to ED physicians include only the inconvenience of completing a survey 2 times over 5 years, the inconvenience of participating in continuing medical education sessions
and the possible disclosure of their knowledge and attitudes toward this treatment. Measures are in place to minimize risk. Surveys can be completed at any time convenient to the physician, CME activities are optional, kept to a minimum of time and offered in a variety of modes (in person, CD, video and web), and all physician-specific data coded to protect confidentiality and carefully guarded against inadvertent disclosure.
Patients at the participating hospitals who are treated with tPA are considered to be secondary subjects as their treatment and the outcomes of that treatment represent the endpoints of the study and will be determined by retrospective review of their medical records. No experimental therapy is administered to patients. As such, the primary risk is the inadvertent disclosure of health information. Multiple measures are employed to minimize this potential risk.
Benefits of the Study:
Physician participants and other ED staff will develop improved knowledge of acute stroke care. Educational and support materials will remain available to staff after the study is completed. Results of this study will be broadly applicable to ED physicians practicing in Michigan and other areas of the country. Increasing the proportion of acute ischemic stroke patients treated appropriately with tPA can be expected to increase the number of stroke patients who recover with minimal to no residual deficits, improving their quality of life and reducing health care expenditures. It is possible that a reduction in stroke morbidity could reduce health care expenditures and increase the contributions of recovered stroke patients to society.
Protection of PHI and Data Storage:
Physician identifications are kept in password protected computer files and no identifiers were used during focus groups or individual interviews. Retrospective charts/images are de-identified and assigned a case number and only the case number appears on data sent outside of the study site. Patient identifiers with the exception of date of birth and treatment dates are not collected. The data being collected meets the definition of a “Limited Data Set”, and therefore, a Data Use Agreement between the University of Michigan and the clinical site has been executed to allow access to this information in lieu of individual patient authorization. Hospital-specific information is identified only by site codes, which are kept confidential from other sites. All written records at the investigational site and at the Clinical Coordinating Center are kept in locked file cabinets inside of locked offices. The electronic data kept by the data management core is kept in password protected files. No individual patient, physician, or hospital data will be disclosed at anytime during the study or after the study is completed. Only aggregate data will be included in any presentations or publications.
Informed Consent:
1. Physicians review information about the study and are invited to participate as part of the on-line survey. They are asked to “accept” (electronically acknowledge the information) before proceeding with the survey. This process occurs each time the survey is administered over the course of the study. The process has been approved as “waiver of documentation of informed consent”.
2. A written informed consent was obtained in person from any individuals who agreed to participate in any on-site focus groups or guided interviews (intervention sites only).
3. The INSTINCT study has been granted a waiver of informed consent for the retrospective review. This study does not involve any contact with the patient or family, nor does it involve any alteration in their clinical care. Stroke patients being treated with tPA are critically ill and often have impaired communication skills and mental acuity. To obtain consent in person would be an unnecessary intrusion and could interfere with or delay important clinical treatment. Most patients will have been discharged and some may expire prior to becoming known to the site investigator. The study does not provide financial resources for on-site personnel to attempt to obtain written informed consent for a retrospective review.
Conclusion:
Finally, stroke treatment with tPA has been approved by the FDA in 1996 and is considered acceptable practice. Thus, the known risks of treatment are not considered risks directly associated with this study. In previously published data, educational efforts aimed at improving physician performance in treating stroke patients with tPA has demonstrated improved performance and reduced risk to patients.
