The PCORnet Bariatric Study: Comparing effectiveness of the 3 most - PowerPoint PPT Presentation

The PCORnet Bariatric Study: Comparing effectiveness of the 3 most common weight loss procedures Neely Williams, MDiv Kathleen M. McTigue, MD MPH Associate Professor of Medicine, Community co-Principal Investigator, Mid-South Clinical Epidemiology & Clinical/Translational Research Network & Meharry-Vanderbilt Alliance Science, University of Pittsburgh Community Partner #PCORI2018

Kathleen McTigue Neely Williams • Have nothing to disclose. 2 • November 17, 2018

Acknowledgements PCORI • Core Scientific Team • • David Arterburn: Lead PI, clinician researcher • Kathleen McTigue: Co-PI, clinician researcher • Neely Williams: Co-PI, patient partner • Karen Coleman: Researcher • Cheri Janning: Patient investigator • Anita Courcoulas: Surgeon investigator • Darren Toh: Distributed analyst • Jane Anau: Project manager • Roy Pardee: Informatician • Robert Wellman: Analyst • Yates Coley: Analyst • Andrea Cook: Analyst Stakeholder Advisory Board • PCORnet CDRN & Coordinating Center teams • 3 • November 17, 2018

Objectives At the conclusion of this activity, the participant should be able to: • Describe how the 3 most common US weight loss procedures compare in terms of weight loss, improvement in diabetes risk, & adverse events • Explain how stakeholder engagement contributed to the project • Make more informed decisions about which bariatric procedure may be right for a patient considering weight loss surgery 4 • November 17, 2018

Background • Severe obesity is a serious health concern affecting 7.7% of Americans (SG) • Use of bariatric surgery has expanded considerably • Sleeve gastrectomy procedure has been used increasingly over past decade – despite a lack of data comparing its effectiveness to other procedures • PCORnet provided a unique opportunity to use real-world health data from 45 health systems to compare bariatric procedures 5 • November 17, 2018

How did stakeholders contribute to the research idea? • PCORnet Obesity Task Force (2014) – Patients, clinicians & researchers prioritized obesity research topics. PCORI then released a funding announcement focusing on weight loss surgery. • At PBS kick-off meeting, requested two major changes to science: • Do three pair-wise comparisons of bariatric procedures, as opposed to two pair-wise comparisons • Interview bariatric surgeons as part of qualitative aim – not just conduct patient focus groups Study activities were carried out, with stakeholder input on data collection & interpretation. 6 • November 17, 2018

How did our stakeholders help us develop & execute our scientific aims? • Reviewed plans to identify cohort. • Includes reviewing diabetes medication lists and bariatric surgery procedure codes. • Using the same process as investigators, prioritized analyses of patient sub-groups. Final rankings were decided by investigators & stakeholders. • Actively participated in development of focus group & surgeon interview templates. 7 • November 17, 2018

Among 46,510 patients from 22 states whose data contributed to weight loss analyses… • Most procedures were bypass or sleeve • 24,982 RYGB (53%) • 18,961 SG (41%) • 2,567 AGB (6%) • The sample was, on average, middle-aged, mostly female & fairly racially/ethnically diverse • Mean age 46; 80% female; 21% Hispanic; 21% African American • Patients were severely obese • Mean BMI: 49 kg/m 2 with 38% BMI 50+ kg/m 2 • Comorbidities were common • 60% HTN; 49% Dyslipidemia; 49% OSA; 40% GERD; 37% T2DM 8 • November 17, 2018

SG & RYGB led to substantial weight loss, which reached a nadir by 1.5 years of follow-up. AGB was less effective. Weight Regain 3.6 kg ↑ AGB 8.2 kg ↑ SG 7.6 kg ↑ RYGB 9 • November 17, 2018

RYGB & SG showed similar DM Remission rates; both led to more remission than AGB Adjusted cumulative remission (%n) RYGB SG AGB 0 1 2 3 4 5 Years since surgery 10 • November 17, 2018

Cumulative relapse was ~67% lower for RYGB (HR 0.32) or SG (HR 0.33) patients compared with AGB patients Adjusted cumulative remission (%n) RYGB AGB SG AGB SG AGB SG RYGB RYGB 0 1 2 3 4 5 0 1 2 3 4 5 Years since surgery 11 • November 17, 2018

RYGB patients had 25% lower relapse rate than SG patients (HR 0.75, 95% CI: 0.67-0.84) Adjusted cumulative remission (%n) RYGB AGB SG AGB SG AGB SG RYGB RYGB 0 1 2 3 4 5 0 1 2 3 4 5 Years since surgery 12 • November 17, 2018

RYGB had the largest & most persistent impact on HbA1c Adjusted HbA1c ∆ from Baseline ∆ HbA1c AGB -0.42% (-0.78, -0.05) -0.45% (-0.63, -0.27) SG RYGB 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Years since surgery 13 • November 17, 2018

Adverse event analyses were restricted to a subset of sites: • Health systems with existing linkages to insurance claims & death data or sites with sufficient samples & ability to link to claims & death data • 34,089 adults from 10 sites in 5 CDRNs • Focus here on 5-year outcomes: • Reoperation • Reoperation with endoscopy • Rehospitalization • Mortality 14 • November 17, 2018

Adjusted 5-year Adverse Events SG RYGB AGB blank (n=15504) (n=18056) (n=1154) Any reoperation 18% 20% 28% Any reoperation or endoscopy 23% 30% 31% Rehospitalization (all cause) 33% 38% 42% Death (all cause) 0.84% 0.89% 1.08% 15 • November 17, 2018

Limitations • Confounding that may have persisted despite covariate & propensity score adjustment • Missing BMI, HbA1c data may introduce bias • Sensitivity analyses suggest missing data were unlikely to change the interpretation of our main results • Comorbid health conditions identified from ICD-9 may underestimate prevalence, can be inaccurately coded, & do not account for severity • AGB procedure under-represented as often carried out in small ambulatory surgical centers • DM medication use estimated from prescribing data, not dispensing, & does not account for adherence 16 • November 17, 2018

Conclusions • Among patients from diverse clinical & geographical settings, SG & RYGB led to substantial weight loss & improvements in diabetes • Plus superior weight loss & diabetes outcomes, and less reoperation, reintervention and rehospitalization than AGB • Compared with SG, RYGB resulted in: • Moderate additional weight loss • Overall, similar rates of DM remission • Larger & more persistent improvement in glycemic control • Lower rates of diabetes relapse • More reoperation, reintervention, & rehospitalization 17 • November 17, 2018

Broader Implications • PCORnet allows for efficient research in a large, geographically- & racially-diverse population accessing academic & community health care settings • Stakeholder engagement ensured clinical relevance • The data enabled timely comparisons of the effectiveness of common procedures that: • Can help patients & clinicians make clinical decisions that best reflect the individual’s risks & benefits • Can help payers understand when a particular procedure may/may not hold particular benefit 18 • November 17, 2018

Learn More • www.pcori.org • info@pcori.org • #PCORI2018 19 • November 17, 2018

Questions? 20 • November 17, 2018

Thank You! Neely Williams, MDiv Kathleen M. McTigue, MD MPH Associate Professor of Medicine, Community co-Principal Investigator, Mid-South Epidemiology & Clinical/Translational Clinical Research Network & Meharry-Vanderbilt Science, University of Pittsburgh Alliance Community Partner 21 • November 17, 2018