Creation of a Translational Science Platform to Foster a Learning Health Care System within the Vanderbilt University Medical Center Enterprise

Transcription

1 Creation of a Translational Science Platform to Foster a Learning Health Care System within the Vanderbilt University Medical Center Enterprise CTSA/VICTR Initiative *********************** PCC and CEG Engagement ***********************

2 Interaction and Information Flow Goal Working toward Enhanced Integration Operations Research LHS Potential

3 CTSA Learning Health Care System Platform Intervention 2 Intervention 1 Studies run through reusable platform Improved Outcomes? NO YES Intervention n Scientific Review LHS Council TRANSLATIONAL SCIENCE PLATFORM Publish De-implement Implement & Publish STUDY SPECIFIC TAILORING Pragmatic Research Studio protocol customization endpoint selection CREATE REUSABLE FIXED INFRASTRUCTURE institutional and regulatory policy, master protocols clinical decision support, clinical registries real-time EHR data capture with coupled REDCap support expert cores, studios, standing DSMB

4 How do we achieve this at VUMC? CTSA Priority Score: 12 Achieves grant funding for all supportive personl effort (e.g. coordination, evidence synthesis, regulatory) promote discovery as routi course in health care delivery encourage innovation and ideas from all clinical areas support and develop research endeavors to answer real world questions enhance quality and safety embrace change implementation

5 Informed Consent Waiver of Informed Consent considered when: Minimal Risk Waiver does not adversely affect the rights or welfare of the participants Consent Impracticable When appropriate, participants will be provided with additional pertint information after participation

6 Learning System in Action at VUMC Meaningful research can be operationally OVERLAID And without delaying implementation

7 Design Design: cluster randomized, multiple crossover trial Setting: five VUMC ICUs from July 2012 through July 2013 Participants: Inclusion: all patients admitted to participating ICUs Exclusion: burns, TEN/SJS, known chlorhexidi allergy IRB approval with waiver of consent

8 Intervention All patients bathed daily with no-rinse cloths: impregnated with chlorhexidi without chlorhexidi All other infection control procedures were performed according to usual practice of each ICU for the duration of the study

9 Prespecified primary outcome Composite rate of healthcare-associated infections CLABSI CAUTI C. difficile infection VAP All outcomes used existing hospital mechanisms for data collection Infection-related outcomes adjudicated by infection control personl Blinded to treatment assignment Used CDC-NHSN definitions

10 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

11 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

12 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

13 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

14 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

15 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

16 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria 10-weeks intervention n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

17 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria 2-week washout n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

18 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

19 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

20 Cardiovascu lar Medical Neuro Surgical Trauma 10,783 patients admitted Units randomized to bathing sequence 0 met exclusion criteria n=1892 n=2327 n=1723 n=1272 n=2126 n=4448 9,340 intention to treat population 1443 admitted during washout excluded n=4852

21 MJ Noto and coauthors Bathing and Health Care Associated Infections: A Randomized Clinical Trial Published onli January 20, 2015 Available at jama.com and on The JAMA Network Reader at mobile.jamatwork.com jamatwork.com

22 Conclusion These findings do not support daily bathing of critically ill patients with chlorhexidi.

23

24 The Challenge: Enroll >5,000 ICU patients delivery of a time-sensitive intervention Collect patient-level fluid, lab, outcome data

25 The Challenge: Enroll >5,000 ICU patients delivery of a time-sensitive intervention Collect patient-level fluid, lab, outcome data Traditional Approach: - Half a decade - Hundreds of personl - Dozens of centers - Millions of dollars

26 The Challenge: Enroll >5,000 ICU patients delivery of a time-sensitive intervention Collect patient-level fluid, lab, outcome data Traditional Approach: - Half a decade - Hundreds of personl - Dozens of centers - Millions of dollars Alternative Approach: - Novel study structure - Informatics-enhanced: - Intervention - Data collection

27 Study Structure Jun Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr MICU X Y X Y X Y X Y X Y X Y Cluster-randomized Neuro Y X Y X Y X Y X Y X Y X CVICU Y X Y X Y X Y X Y X Y X each ICU randomized to a Fluid Group ( NS vs. balanced ) Trauma Y X Y X Y X Y X Y X Y X Multiple-crossover SICU Y X Y X Y X Y X Y X Y X 1 month blocks, 11 crossovers (o year) 8,000 patients ED Y X Y X Y X Y X Y X Y X Y X

28 Study Structure X = Normal sali, Y = balanced (Plasmalyte or LR) Jun Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr MICU X Y X Y X Y X Y X Y X Y Cluster-randomized Neuro Y X Y X Y X Y X Y X Y X CVICU Y X Y X Y X Y X Y X Y X each ICU randomized to a Fluid Group ( NS vs. balanced ) Trauma Y X Y X Y X Y X Y X Y X Multiple-crossover SICU Y X Y X Y X Y X Y X Y X 1 month blocks, 11 crossovers (o year) 8,000 patients ED Y X Y X Y X Y X Y X Y X Y X

29 Delivery of the Intervention Pharmacy stocks unit preferentially with assigd fluid

30 Delivery of the Intervention Pharmacy stocks unit preferentially with assigd fluid

31 Delivery of the Intervention If assigd to Normal Sali, can bypass for: - specific attending request If assigd to Balanced Fluid can bypass for: - hyperkalemia - brain injury - specific attending request Pharmacy stocks unit preferentially with assigd fluid

32 MEASUREMENTS AND MAIN RESULTS: Patients assigd to sali (n=454) and balanced crystalloids (n=520) were similar at baseli and received similar volumes of crystalloid by 30 days (2,715 ± 4,073 ml vs 3,157 ± 5,069 ml; P = 0.40). MAKE30 did not differ between groups (24.7% vs 24.6%; P = 0.98).

33 Video vs. Direct Laryngoscopy in Critically Ill Patients Patient Level Randomization LHS Project desigd by Pulm/CC traie

34 Study Design (Fellow Project) Open-label, parallel-group, randomized trial

35 Study Design (Fellow Project) Open-label, parallel-group, randomized trial Video Laryngoscopy vs. Direct Laryngoscopy

36 Study Design (Fellow Project) Open-label, parallel-group, randomized trial Video Laryngoscopy vs. Direct Laryngoscopy Inclusion: 18 or older being intubated by PCCM fellow

37 Study Design (Fellow Project) Open-label, parallel-group, randomized trial Video Laryngoscopy vs. Direct Laryngoscopy Inclusion: 18 or older being intubated by PCCM fellow Exclusion: awake intubation too emergent to open study envelope specific laryngoscope required

38 Video Laryngoscopy Improves Glottic Visualization

39 Video Laryngoscopy did not Increase First Attempt Success

40 Video Laryngoscopy Does Not Decrease Time to Intubation

41 Crit Care Med Nov;44(11):

42 A Pragmatic, Randomized, led Trial Examining the Efficacy of a Hospital Discharge Follow-up Pho Call Program Maya Yiadom, MD Principle Investigator

43 Study Procedure Cohort Identification Stephen Pert Daily busiss objects reports every weekday morning for ~6-9 months All inpatient status hospital discharges that occurred the previous 24 hrs (or 72hrs on Mondays) Encounter number MRN Admit date Discharge date Name All pho numbers available Street address City, state Busiss Intelligence Analyst Zipcode Department of Finance Sent, every week day, to a Randomization Nurse via RedCap

44 Study Procedure Intervention Delivery/Exposure Pho Call Program Pho Call Nurse will: Place a call within 3 days of discharge Repeat calls made (3 calls max) for up to 7 days post discharge Make call guided by semi-structured script Screen for w symptoms Review understanding of discharge meds Review understanding of follow-up appt Identify ed for intervention Interventions provided Document Call outcomes in Starform Reasons patients are unreached Call attempts Sarah Marlow, RN Operations Systems Engier Discharge Pho Call Program

45 Study Procedure Post Enrollment Covariate Data Collection Research Derivative Team CCSIR will pull covariates Starpal Data Dumps (VICTR Randomization RN) Statitsical Analysis Q1month Q1week Dikshya Interim Bastakoty Emily Bruer Descriptive, Univariate and Multivariate Kevin Cox VICTR Data Analyst Translational Research Data Coordinator Research Derivative Analyses Research Derivative Final Descriptive, Univariate and Multivariate Analyses

46 Post Enrollment Covariate Data Collection Research Derivative Team CCSIR will pull covariates Starpal Data Dumps Q1week (VICTR Randomization RN) Statitsical Analysis Interim Descriptive, Univariate and Multivariate Analyses Final Descriptive, Univariate and Multivariate Analyses Sunil Kripalani, MD Director, Center for Clinical Implementation Sciences Research (CCISR) Kathryn Goggins Research Program Manager, Center for Clinical Implementation Sciences Research (CCISR)

47 REDCap Data Base Management Build study REDCap Database Receive data pulls from, Finance, RD and CCQIR Import data from Starpal data dumps Alert team to data integrity and quality issues that may produce bias or missingss Basic level data cleaning Export data for biostatistical analysis

48 Interim Analysis Interim Analysis Safety Officer: Tina Hartert, MD Asst. Vice President for Translational Science Blinded Statistician: Li Wang, MS VICTR Biostatistician II Blinded Results Review: Select VUMC Executives Mitch Edgeworth, MBA Chief Executive Officer Jerry Hickson, MD Chief Quality, Patient Safety, and Risk Prevention Officer Robin Steaban, RN Associate Chief Nursing Officer

49 Study Procedure Results Review and Interpretation Results Review and Interpretation: Study Team Maya Yiadom Hank Dominico Dan Byr Michelle Hasselblad Neesha Choma Sunil Kripalani Frank Harrell Sara Marlow Cheryl Gatto Gordon Bernard

50 Results Dissemination VUMC Leadership Adult Enterprise Quality Committee Poster/Abstract Presentation Manuscript publication

51 CTSA Learning Health Care System Platform Intervention 2 Intervention 1 Studies run through reusable platform Improved Outcomes? NO YES Intervention n Scientific Review LHS Council TRANSLATIONAL SCIENCE PLATFORM Publish De-implement Implement & Publish STUDY SPECIFIC TAILORING Pragmatic Research Studio protocol customization endpoint selection CREATE REUSABLE FIXED INFRASTRUCTURE institutional and regulatory policy, master protocols clinical decision support, clinical registries real-time EHR data capture with coupled REDCap support expert cores, studios, standing DSMB

52 Title Idea and Protocol Regulatory IRB/CT Launch and conduct Interim analysis Enrollment end Final analysis and close out SMART ICU Fluids Semler (N=10,000, 24 months)** Discharge Pho Calls Yiadom (N=3556, 6 months) September 2017 AKI Model Peds Van Driest (N=10,000, 12 months) November 2017 COMPASS PC Karlekar/Ely (N=400, 24 months) review after 30 pts enrolled and tracked for 6 months ICU Recovery Program Bloom/Stollings (N=550, 6 months) Procalcitonin PEDS Barjee (N=210, 12 months) Contact Precautions ICU Huerta (N=2000, 8 months) VSP MS Bagwell (in development) ** = estimated enrollment target from ClinicalTrials.gov where registered or study protocol if not yet underway HARD STOP IN PROCESS COMPLETE

53 Questions?