'ppmHR' R package

  

• Download link: CRAN

• Funding source: None

• Description: A R package that allows users to estimate overall and site-specific hazard ratios in multi-database studies using inverse probability weighted Cox proportional hazards model and summary-level riskset tables from data-contributing sites.

• Related publication: Shu D, Yoshida K, Fireman BH, Toh S. Inverse probability weighted Cox model in multi-site studies without sharing individual-level data. Stat Methods Med Res 2020;29(6):1668-1681 [PubMed]

​

​

'distributed' R package for the simulation study in the Privacy-Preserving Analytic and Data-Sharing Methods for Clinical and Patient-Powered Data Networks project

  

• Download link: GitHub

• Funding source: Patient-Centered Outcomes Research Institute (ME-1403-11035)

• Description: A R package that allows users to design and perform simulation studies to compare the performance of different combinations of confounder summarization technique (individual covariates, propensity scores, disease scores), confounding adjustment method (matching, stratification, weighting), and data-sharing approach (individual-level data, summary-table data, risk-set data, and effect-estimate data).​

• Related publication: Yoshida K, Gruber S, Fireman BH, Toh S. Comparison of privacy-protecting analytic and data-sharing methods: a simulation study. Pharmacoepidemiol Drug Saf 2018;27(9):1034-1041 [PubMed]

​

​

SAS code for the empirical study in the Privacy-Preserving Analytic and Data-Sharing Methods for Clinical and Patient-Powered Data Networks project

 

• Download link: GitHub

• Funding source: Patient-Centered Outcomes Research Institute (ME-1403-11035)

• Description: A SAS package that compares the performance of different combinations of confounder summarization technique (individual covariates, propensity scores, disease risk scores), confounding adjustment method (matching, stratification, weighting), and data-sharing approach (individual-level data, summary-table data, risk-set data, and effect-estimate data) using two empirical examples (comparative effectiveness and safety of bariatric procedures; comparative effectiveness and safety of biologic treatments for rheumatoid arthritis).

• Related publication: Li X, Fireman BH, Curtis JR, Arterburn DE, Fisher DP, Moyneur E, Gallagher M, Raebel MA, Nowell WB, Lagreid L, Toh S. Validity of privacy-protecting analytical methods that use only aggregate-level information to conduct multivariable-adjusted analysis in distributed data networks. Am J Epidemiol 2019;188(4):709-723 [PubMed]

​

​

SAS code for distributed linear, logistic, and Cox regression

 

• Download link: GitHub or Sentinel Initiative

• Funding source: Office of the Assistant Secretary for Planning and Evaluation & Food and Drug Administration (HHSF223201400030I/HHSF22301006T)

• Description: Two SAS packages that allow users to perform multivariable-adjusted distributed linear, logistic, and Cox regression in a multi-database study. One package, to be executed by participating data partners, produces the necessary summary-level information (e.g., sums of squares and cross products matrix) to be sent to the analysis center. Another package, to be executed by the analysis center, aggregates the summary-level information from participating sites and produces the final results. The SAS packages are fully integrated with PopMedNet, a distributed networking software platform, to allow manual, semi-automated, and fully automated distributed regression analysis.

• Related publication: Her QL, Malenfant JM, Malek S, Vilk Y, Young J, Li L, Brown J, Toh S. A query workflow design to perform automatable distributed regression analysis in large distributed data networks. eGEMs: The Journal of Electronic Health Data and Methods 2018;6(2):11 [PubMed​][Free Fulltext]

​

​

SAS code for distributed linear regression combined with propensity scores in the PCORnet Bariatric Study

 

• Download link: GitHub

• Funding source: Patient-Centered Outcomes Research Institute (OBS-1505-30683)

• Description: Two SAS packages that allow users to perform multivariable-adjusted distributed linear regression in a multi-database study. One package, to be executed by participating data partners, produces the necessary summary-level information (e.g., sums of squares and cross products matrix) to be sent to the analysis center. Another package, to be executed by the analysis center, aggregates the summary-level information from participating sites and produces the final results. The analysis combines propensity scores with distributed linear regression, providing doubly privacy protection. 

• Related publication: Toh S, Wellman R, Coley RY, Horgan C, Sturtevant J, Moyneur E, Janning C, Pardee R, Coleman KJ, Arterburn D, McTigue K, Anau J, Cook AJ, on behalf of the PCORnet Bariatric Study Collaborative. Combining distributed regression and propensity scores: A doubly privacy-protecting analytic method for multicenter research. Clin Epidemiol 2018;10:1773-1786 [PubMed][Free Fulltext]

​

​

SAS code for distributed linear regression in the PCORnet Antibiotics and Childhood Growth Study

 

• Download link: GitHub

• Funding source: Patient-Centered Outcomes Research Institute (OBS-1505-30699)

• Description: Two SAS packages that allow researchers to perform multivariable-adjusted distributed linear regression in a multi-database study. One package, to be executed by participating data partners, produces the necessary summary-level information (e.g., sums of squares and cross products matrix) to be sent to the analysis center. Another package, to be executed by the analysis center, aggregates the summary-level information from participating sites and produces the final results. 

• Related publication: Forthcoming

​

​

Sentinel Routine Querying System

 

• Download link: Sentinel Initiative

• Funding source: Food and Drug Administration (HHSF223201400030I)

• Description: A suite of SAS programs developed to be executed against the Sentinel Common Data Model. One of the modules allow users to perform propensity score matching or stratification using only risk set data. 

• Related publication: Gagne JJ, Han X, Hennessy S, Leonard CE, Chrischilles EA, Carnahan RM, Wang SV, Fuller C, Iyer A, Katcoff H, Woodworth TS, Archdeacon P, Meyer TE, Schneeweiss, Toh S. Successful comparison of US Food and Drug Administration Sentinel analysis tools to traditional approaches in quantifying a known drug-adverse-event association. Clin Pharmacol Ther 2016;100(5):558-564 [PubMed]

​

​

​