TY - JOUR
T1 - Cross-design synthesis for extending the applicability of trial evidence when treatment effect is heterogenous
T2 - Part II. Application and external validation
AU - Henderson, Nicholas C.
AU - Varadhan, Ravi
AU - Weiss, Carlos O.
N1 - Funding Information:
The authors would like to acknowledge the support of Dr. Parivash Nourjah, project officer, AHRQ. Dr. Varadhan was also supported by a Brookdale Leadership in Aging Fellowship. The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the US Department of Health and Human Services. This work used research materials obtained from the National Heart Lung and Blood Institute (NHLBI) Biologic Specimen and Data Repository Information Coordinating Center and does not necessarily reflect the opinions or views of the Studies of Left Ventricular Dysfunction or the NHLBI. The authors also thank Drs. Jodi Segal, Cynthia Boyd, and Albert Wu of Johns Hopkins University and Dr. David Kent of Tufts University for their valuable feedback during the conduct of this research.
Funding Information:
Agency for Healthcare Research and Quality, US Department of Health and Human Services (HHSA29020050034-I-TO5) The authors would like to acknowledge the support of Dr. Parivash Nourjah, project officer, AHRQ. Dr. Varadhan was also supported by a Brookdale Leadership in Aging Fellowship. The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the US Department of Health and Human Services. This work used research materials obtained from the National Heart Lung and Blood Institute (NHLBI) Biologic Specimen and Data Repository Information Coordinating Center and does not necessarily reflect the opinions or views of the Studies of Left Ventricular Dysfunction or the NHLBI. The authors also thank Drs. Jodi Segal, Cynthia Boyd, and Albert Wu of Johns Hopkins University and Dr. David Kent of Tufts University for their valuable feedback during the conduct of this research.
Publisher Copyright:
© 2017, © 2017 Taylor & Francis.
PY - 2017/4/3
Y1 - 2017/4/3
N2 - Randomized controlled trials (RCTs) generally provide the most reliable evidence. When participants in RCTs are selected with respect to characteristics that are potential treatment effect modifiers, the average treatment effect from the trials may not be applicable to a specific target population. We present an application of the recently developed calibrated risk-adjusted modeling (CRAM) method for projecting the treatment effect from an RCT to a target group that is inadequately represented in the trial when there is heterogeneity in the treatment effect (HTE). The CRAM method allows for integration of RCT and observational data through cross-design synthesis. An essential component of CRAM is to identify HTE and to then compute a calibration factor for unmeasured confounding for the observational study relative to the RCT. The estimate of treatment effect adjusted for unmeasured confounding is projected onto the target sample using G-computation with standardization weights. In this paper, we apply CRAM to estimate the effect of angiotensin converting enzyme inhibition to prevent heart failure hospitalization or death. External validation shows that when there is adequate overlap between the RCT and the target sample, risk-based standardization is less biased than CRAM. However, when there is poor overlap between the trial and the target sample, CRAM provides superior estimates of treatment effect.
AB - Randomized controlled trials (RCTs) generally provide the most reliable evidence. When participants in RCTs are selected with respect to characteristics that are potential treatment effect modifiers, the average treatment effect from the trials may not be applicable to a specific target population. We present an application of the recently developed calibrated risk-adjusted modeling (CRAM) method for projecting the treatment effect from an RCT to a target group that is inadequately represented in the trial when there is heterogeneity in the treatment effect (HTE). The CRAM method allows for integration of RCT and observational data through cross-design synthesis. An essential component of CRAM is to identify HTE and to then compute a calibration factor for unmeasured confounding for the observational study relative to the RCT. The estimate of treatment effect adjusted for unmeasured confounding is projected onto the target sample using G-computation with standardization weights. In this paper, we apply CRAM to estimate the effect of angiotensin converting enzyme inhibition to prevent heart failure hospitalization or death. External validation shows that when there is adequate overlap between the RCT and the target sample, risk-based standardization is less biased than CRAM. However, when there is poor overlap between the trial and the target sample, CRAM provides superior estimates of treatment effect.
KW - Generalizability
KW - heterogeneity
KW - interaction
KW - internal and external validity
KW - observational data
KW - real-world evidence
KW - sensitivity analysis
KW - unmeasured confounding
UR - http://www.scopus.com/inward/record.url?scp=85064420633&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85064420633&partnerID=8YFLogxK
U2 - 10.1080/23737484.2017.1398056
DO - 10.1080/23737484.2017.1398056
M3 - Article
AN - SCOPUS:85064420633
SN - 2373-7484
VL - 3
SP - 7
EP - 20
JO - Communications in Statistics Case Studies Data Analysis and Applications
JF - Communications in Statistics Case Studies Data Analysis and Applications
IS - 1-2
ER -