TY - JOUR
T1 - Privacy-protecting multivariable-adjusted distributed regression analysis for multi-center pediatric study
AU - on behalf of the PCORnet Antibiotics and Childhood Growth Study Group
AU - Toh, Sengwee
AU - Rifas-Shiman, Sheryl L.
AU - Lin, Pi I.D.
AU - Bailey, L. Charles
AU - Forrest, Christopher B.
AU - Horgan, Casie E.
AU - Lunsford, Douglas
AU - Moyneur, Erick
AU - Sturtevant, Jessica L.
AU - Young, Jessica G.
AU - Block, Jason P.
AU - Appelhans, Brad
AU - Arterburn, David
AU - Boone-Heinenon, Janne
AU - Brickman, Andrew L.
AU - Bunnell, H. Timothy
AU - Cole, F. Sessions
AU - Daley, Matthew F.
AU - Dempsey, Amanda
AU - Finkelstein, Jonathan
AU - Fitzpatrick, Stephanie L.
AU - Heerman, William
AU - Horberg, Michael
AU - Isasi, Carmen R.
AU - Jay, Melanie
AU - Kharbanda, Elyse
AU - Khare, Ritu
AU - Lemas, Dominick
AU - Lin, Simon M.
AU - Messito, Mary Jo
AU - O’Neill, Allison
AU - Peay, Holly Landrum
AU - Prochaska, Micah
AU - Ranade, Daksha
AU - Rao, Goutham
AU - Rayas, Maria
AU - Reynolds, Juliane S.
AU - Rosenman, Marc
AU - Taylor, Bradley
AU - Willis, Zachary
N1 - Publisher Copyright:
© 2019, International Pediatric Research Foundation, Inc.
PY - 2020/5/1
Y1 - 2020/5/1
N2 - Background: Privacy-protecting analytic approaches without centralized pooling of individual-level data, such as distributed regression, are particularly important for vulnerable populations, such as children, but these methods have not yet been tested in multi-center pediatric studies. Methods: Using the electronic health data from 34 healthcare institutions in the National Patient-Centered Clinical Research Network (PCORnet), we fit 12 multivariable-adjusted linear regression models to assess the associations of antibiotic use <24 months of age with body mass index z-score at 48 to <72 months of age. We ran these models using pooled individual-level data and conventional multivariable-adjusted regression (reference method), as well as using the more privacy-protecting pooled summary-level intermediate statistics and distributed regression technique. We compared the results from these two methods. Results: Pooled individual-level and distributed linear regression analyses produced virtually identical parameter estimates and standard errors. Across all 12 models, the maximum difference in any of the parameter estimates or standard errors was 4.4833 × 10−10. Conclusions: We demonstrated empirically the feasibility and validity of distributed linear regression analysis using only summary-level information within a large multi-center study of children. This approach could enable expanded opportunities for multi-center pediatric research, especially when sharing of granular individual-level data is challenging.
AB - Background: Privacy-protecting analytic approaches without centralized pooling of individual-level data, such as distributed regression, are particularly important for vulnerable populations, such as children, but these methods have not yet been tested in multi-center pediatric studies. Methods: Using the electronic health data from 34 healthcare institutions in the National Patient-Centered Clinical Research Network (PCORnet), we fit 12 multivariable-adjusted linear regression models to assess the associations of antibiotic use <24 months of age with body mass index z-score at 48 to <72 months of age. We ran these models using pooled individual-level data and conventional multivariable-adjusted regression (reference method), as well as using the more privacy-protecting pooled summary-level intermediate statistics and distributed regression technique. We compared the results from these two methods. Results: Pooled individual-level and distributed linear regression analyses produced virtually identical parameter estimates and standard errors. Across all 12 models, the maximum difference in any of the parameter estimates or standard errors was 4.4833 × 10−10. Conclusions: We demonstrated empirically the feasibility and validity of distributed linear regression analysis using only summary-level information within a large multi-center study of children. This approach could enable expanded opportunities for multi-center pediatric research, especially when sharing of granular individual-level data is challenging.
UR - http://www.scopus.com/inward/record.url?scp=85074628913&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85074628913&partnerID=8YFLogxK
U2 - 10.1038/s41390-019-0596-0
DO - 10.1038/s41390-019-0596-0
M3 - Article
C2 - 31578038
AN - SCOPUS:85074628913
SN - 0031-3998
VL - 87
SP - 1086
EP - 1092
JO - Pediatric Research
JF - Pediatric Research
IS - 6
ER -