12/30/2011. Dan Spaite : PI NIH/NINDS 1R01NS A1. Ben Bobrow: PI NIH/NINDS 1R01NS A1

Transcription

1 Daniel Spaite, MD, FACEP Professor of Emergency Medicine Ben Bobrow, MD, FACEP Associate Professor of Emergency Medicine Dan Spaite : PI NIH/NINDS 1R01NS A1 Ben Bobrow: PI NIH/NINDS 1R01NS A1 Main objectives of this session: Present the scientific foundation and issues of implementation in the EPIC TBI Project Assist those aspiring to win federal grants for EMS research Approach to this session: Use the new NIH R01 Grant application format Three NIH Project Plan sections: Significance Innovation Approach 1

2 TBI is the leading cause of death and disability worldwide USA 5.3 million Americans require long term assistance with daily activities due to TBI 2% of the population DIRECT cost = 60 billion/year (2000) Significance: TBI in the US: ,000 Deaths 235,000 Hospitalizations 1,111,000 Emergency Department Visits Receiving Other Medical Care or No Care? At least 1.4 million TBIs occur in the U.S. each year.* * Average annual numbers, Significance: TBI in the US: 2010 ~1.7 million total TBIs 52,000 Deaths 275,000 Hospitalizations Up 4% 1,365,000 Emergency Department Visits Receiving Other Medical Care or No Care? Up 17% Up 23% Faul: TBI in the US; CDC Natl Ctr for Injury Prevention and Control;

3 What happens in the first few minutes profoundly impacts outcome The Killer H s for TBI Hypoxemia Hypotension Hyperventilation Hypoxia: Hypoxia in the field is very common Occurs in as many as 55% of patients with severe TBI A single non-spurious O 2 sat of <90% is independently associated with at least a doubling of mortality One study: Tripled mortality Hypotension: A single episode of SBP <90 is independently associated with at least a doubling of mortality Repeated episodes (OR for death = 8) 3

4 Hyperventilation (intubated pts): Hyperventilation is independently associated with at least a doubling of mortality One study showed a six-fold increase in mortality for patients with hyperventilation at any time during EMS/early hospital care Inadvertent hyperventilation is automatic if not meticulously prevented: NO ONE (esp in EMS/ED) can ventilate properly without adjuncts (ETCO 2 monitoring, cadence devices, ventilators) This substantial body of evidence has led to nationally-vetted EBGs The essence of these Guidelines: Prevent and aggressively treat hypoxia Prevent and aggressively treat hypotension Meticulously prevent and rapidly correct hyperventilation in intubated patients 4

5 Developed or Endorsed by: The Brain Trauma Foundation National Highway Traffic Safety Administration National Association of EMS Physicians Many others 2 nd edition published in Prehospital Emergency Care in State EMS Med Directors surveyed: Essentially no EMS agencies have implemented any of the guidelines: (65%) Only one of these states was aware of any EMS agencies planning on implementing in the future Only a few agencies have implemented (20%) A substantial number of agencies have implemented (15%) None have evaluated whether guideline treatment: Is actually being performed Implementation has impacted outcomes EMS in Arizona: Since 2004 Establishment of a partnership to improve outcomes: >100 EMS agencies (>80% of the population) with Arizona DOH & University of Arizona The collaboration has carried out: An integrated plan of implementing and measuring improvements in EMS care across the state Linkage of EMS data to distal outcomes The ability to analyze the impact of these changes 5

6 Widespread implementation of bystander Compression-only CPR (COCPR) and increased CPR rates Implementation of EMS Minimally-Interrupted Cardiac Resuscitation (MICR) associated with improved outcomes Implementation of a network of specialized Cardiac Receiving Centers (CRCs) and EMS bypass protocols for cardiac arrest patients with ROSC Bystander interventions for OHCA Statewide campaign to: Increase bystander CPR rates Officially endorse and encourage COCPR BCPR measurement 100% 80% 60% 40% 20% 0% 77% % Lay COCPR 481% RI 41.5% RI 28.2% 39.9% All Lay CPR P = % Bobrow, Spaite, Berg, et al: JAMA;2010 6

7 Implementation and measurement of EMS Minimally-Interrupted Cardiac Resuscitation (MICR) Survival to Hospital Discharge (%) MICR ACLS (61/1686) 3.6 (55/598) 9.2 aor = 3.0 (38/348) 10.9 (36/128) 28.1 All cardiac arrests Witnessed with VF Bobrow, et al. JAMA 2008 Vol. 299 No. 10 Implementation of a network of specialized Cardiac Receiving Centers (CRCs) and EMS bypass protocols for post cardiac arrest patients with ROSC 7

8 % Survival OR: 2.15 [ ] OR: 2.03 [ ] OR:2.00 [ ] Shockable Non-shockable Witnessed VF Pre Post Spaite et al. in press This partnership has created the environment where the evaluation of the impact of implementing the TBI guidelines can occur Statewide implementation and evaluation of the National EMS TBI Guidelines 8

9 Aim #1: Implement the nationallyvetted TBI Guidelines across a vast network of EMS systems in Arizona >100 EMS agencies Urban, suburban, rural, frontier Public and private Ground and air EMS >80% of the population Aim #2: Collect and link prehospital and trauma center data on moderate and severe TBI patients to determine: Injury severity/risk-adjustment measures Treatment Outcomes Done in the pre-implementation and post-implementation cohorts Aim #3: Evaluate the impact of prehospital Guideline implementation on multiple outcomes in patients with moderate and severe TBI 9

10 Phase 1: 5 years of retrospective baseline data EMS data: Prehospital PCRs Outcomes: Arizona State Trauma Registry Run-in Phase: The time in each agency from the beginning of training until training is complete and Guidelines implemented Phase 2: 4 years of prospective, postimplementation data Phase I Phase I (R) Phase II (baseline: mostly retrospective) (run-in) (prospective: post-intervention) Risk adjustment and outcome measurements established before intervention Begin training Training and implementation complete Analysis Implementation: Train > 10,000 EMS providers via an agencybased train-the-trainer process Flexible educational model to fit individual agency training schedule Training methods: Didactic Interactive, scenario-based training modules Hands-on lab training Video Web-based updates Regular agency feedback 10

11 Estimated number of mod/severe TBI patients: Based upon the ASTR ( ) Over 20,000 total patients Phase 1: ~11,000 Phase 2: ~9,000 Over 2,400 intubated patients Phase 1: ~1,300 Phase 2: ~1,100 Moderate or severe TBI: Anyone with physical trauma and a mechanism consistent with the potential to have a brain injury and: GCS of 12 or less or GCS <15 with decreasing GCS or increasing confusion or Multisystem trauma requiring intubation whether the primary need for intubation was from TBI or from other injuries or Post-traumatic seizures Risk-adjustment/severity measures: VS, GCS, RTS CDC Barell Matrix AIS/ISS TRISS/Probability of Survival Diagnoses: ICD-9 11

12 Outcomes: Primary: Overall Mortality Mortality among intubated patients Secondary Hospital days ICU days Ventilator days Disposition Cost Exempt from HIPAA: State-sponsored Public Health Initiative This allows linkage to outcome data Arizona State Trauma Registry: Allows us to identify each true TBI patient Able to bring the specific cases to the agencies Able to provide specific outcome data to the agencies Baseline data: EMS data and ASTR data from Post-implementation data: Same data for 4 years after guideline implementation Before/after, severity-adjusted comparison of: Risk-adjustment measures EMS treatment rendered Adherence to the Guidelines Treating hypoxia, hypotension, and preventing hyperventilation in intubated patients Outcomes Sequential Analyses Interim analysis: End of year 3 of implementation p < 0.01 Final Analysis: End of year 4 p < 0.04 Sequential Analysis Adjustment Overall α =

13 An EB, public health initiative: Sponsored by the State Department of Health Purpose: Improve patient care and outcomes EMS protocols approved by the State EMS Council and the Director of the ADHS No randomization Voluntary with local jurisdiction autonomy UA and ADHS IRB Regulatory determinations: It is NOT human subjects research This is different than exempt It is exempt from HIPAA Director of ADHS State Attorney General *These two issues are huge for being able to link EMS and hospital outcome data Five year NIH R01 grant (NINDS) The first prehospital observational, before-after system study ever funded by the NIH 13

14 Scoring category: Significance The issue has profound PH implications There are internationally-vetted EBGs that have not been evaluated for their impact on outcome The first 30 minutes of care may dwarf the impact of subsequent care One live brain Scoring category: Approach and Environment: The interventions are simple, discrete, widely available, and feasible The EMS agency/university/state partnership has a track record of: Implementing major changes in EMS care Collecting - linking EMS and hospital data Evaluating the impact of the changes on outcomes Another major hurdle NIH s historical stand: Essentially, they only fund RCTs They don t fund: Observational Research Implementation Research Outcomes Research Comparative Effectiveness Research Thus historically they haven t funded EMS research RAMPART is a rare study 14

15 Argument #1: When bundled care is optimal care there may be no SINGLE component that can be identified as effective in classic RCTs ROC has found: Lack of effect when making a single change in a complex condition TBI may require multiple SIMULTANEOUS interventions to move the needle. Argument #2: RCT not feasible No studies of the prehospital TBI Guidelines However, the supporting evidence is strong Even if one argues that randomization is ethical it is currently not feasible No system would randomization after full disclosure of the evidence E.g., Who would be willing to randomize hyperventilation?? Argument #3: Evaluating large bundles of care is not feasible via a distributive model Research consortia (NETT, ROC) evaluate highly specific clinical questions via randomization Making sweeping changes in EMS is jurisdictional Each state determines protocols Promulgation of widespread changes across states is exceedingly unlikely via a distributive model 15

16 Argument #4: Variations in EBG compliance are inevitable, but, if done correctly, this does not negate the importance of the results This is an effectiveness evaluation deployed in the real world across vast geography/demography This is inherent to the widespread implementation of healthcare interventions (CER), especially in EMS However, variation and granularity will allow stratification of patients for sub-group analysis E.g.: Variation in compliance with managing postintubation ETCO 2 levels Determining the impact of EMS on outcomes Requires: Partnerships EMS -Public Health -Universities E.g., IRB and HIPAA issues Integration of multiple methods -clinical research, outcomes research, systems research, epidemiology Understanding efficacy vs. effectiveness Identifying the best available study design Evaluating complex bundles of care that cannot feasibly be randomized 16

17 ????? Arizona Fire Departments and EMS Agencies 17