Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score
The prognostic score, or disease risk score (DRS), is a summary score that is used to control for confounding in non-experimental studies. While the DRS has been shown to effectively control for measured confounders, unmeasured confounding continues to be a fundamental obstacle in non-experimental r...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
De Gruyter
2014-09-01
|
| Series: | Journal of Causal Inference |
| Subjects: | |
| Online Access: | https://doi.org/10.1515/jci-2014-0009 |
| _version_ | 1819007280875569152 |
|---|---|
| author | Wyss Richard Lunt Mark Brookhart M. Alan Glynn Robert J. Stürmer Til |
| author_facet | Wyss Richard Lunt Mark Brookhart M. Alan Glynn Robert J. Stürmer Til |
| author_sort | Wyss Richard |
| collection | DOAJ |
| description | The prognostic score, or disease risk score (DRS), is a summary score that is used to control for confounding in non-experimental studies. While the DRS has been shown to effectively control for measured confounders, unmeasured confounding continues to be a fundamental obstacle in non-experimental research. Both theory and simulations have shown that in the presence of unmeasured confounding, controlling for variables that affect treatment (both instrumental variables and measured confounders) amplifies the bias caused by unmeasured confounders. In this paper, we use causal diagrams and path analysis to review and illustrate the process of bias amplification. We show that traditional estimation strategies for the DRS do not avoid bias amplification when controlling for predictors of treatment. We then discuss estimation strategies for the DRS that can potentially reduce bias amplification that is caused by controlling both instrumental variables and measured confounders. We show that under certain assumptions, estimating the DRS in populations outside the defined study cohort where treatment has not been introduced, or in outside populations with reduced treatment prevalence, can control for the confounding effects of measured confounders while at the same time reduce bias amplification. |
| first_indexed | 2024-12-21T00:22:05Z |
| format | Article |
| id | doaj.art-ee68d56c763b42c0971270d357708b41 |
| institution | Directory Open Access Journal |
| issn | 2193-3677 2193-3685 |
| language | English |
| last_indexed | 2024-12-21T00:22:05Z |
| publishDate | 2014-09-01 |
| publisher | De Gruyter |
| record_format | Article |
| series | Journal of Causal Inference |
| spelling | doaj.art-ee68d56c763b42c0971270d357708b412022-12-21T19:22:05ZengDe GruyterJournal of Causal Inference2193-36772193-36852014-09-012213114610.1515/jci-2014-0009Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk ScoreWyss Richard0Lunt Mark1Brookhart M. Alan2Glynn Robert J.3Stürmer Til4Department of Epidemiology, UNC Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7435, USAArthritis Research UK Epidemiology Unit, Centre for Musculoskeletal Research, Institute of Inflammation and Repair, Manchester Academic Health Science Centre, The University of Manchester, Manchester, UKDepartment of Epidemiology, UNC Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7435, USADepartment of Biostatistics, Harvard School of Public Health, Brigham & Women’s Hospital, Boston, MA, USADepartment of Epidemiology, UNC Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7435, USAThe prognostic score, or disease risk score (DRS), is a summary score that is used to control for confounding in non-experimental studies. While the DRS has been shown to effectively control for measured confounders, unmeasured confounding continues to be a fundamental obstacle in non-experimental research. Both theory and simulations have shown that in the presence of unmeasured confounding, controlling for variables that affect treatment (both instrumental variables and measured confounders) amplifies the bias caused by unmeasured confounders. In this paper, we use causal diagrams and path analysis to review and illustrate the process of bias amplification. We show that traditional estimation strategies for the DRS do not avoid bias amplification when controlling for predictors of treatment. We then discuss estimation strategies for the DRS that can potentially reduce bias amplification that is caused by controlling both instrumental variables and measured confounders. We show that under certain assumptions, estimating the DRS in populations outside the defined study cohort where treatment has not been introduced, or in outside populations with reduced treatment prevalence, can control for the confounding effects of measured confounders while at the same time reduce bias amplification.https://doi.org/10.1515/jci-2014-0009bias amplificationprognostic scoreunmeasured confoundingpath analysiscausal diagrams |
| spellingShingle | Wyss Richard Lunt Mark Brookhart M. Alan Glynn Robert J. Stürmer Til Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score Journal of Causal Inference bias amplification prognostic score unmeasured confounding path analysis causal diagrams |
| title | Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score |
| title_full | Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score |
| title_fullStr | Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score |
| title_full_unstemmed | Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score |
| title_short | Reducing Bias Amplification in the Presence of Unmeasured Confounding through Out-of-Sample Estimation Strategies for the Disease Risk Score |
| title_sort | reducing bias amplification in the presence of unmeasured confounding through out of sample estimation strategies for the disease risk score |
| topic | bias amplification prognostic score unmeasured confounding path analysis causal diagrams |
| url | https://doi.org/10.1515/jci-2014-0009 |
| work_keys_str_mv | AT wyssrichard reducingbiasamplificationinthepresenceofunmeasuredconfoundingthroughoutofsampleestimationstrategiesforthediseaseriskscore AT luntmark reducingbiasamplificationinthepresenceofunmeasuredconfoundingthroughoutofsampleestimationstrategiesforthediseaseriskscore AT brookhartmalan reducingbiasamplificationinthepresenceofunmeasuredconfoundingthroughoutofsampleestimationstrategiesforthediseaseriskscore AT glynnrobertj reducingbiasamplificationinthepresenceofunmeasuredconfoundingthroughoutofsampleestimationstrategiesforthediseaseriskscore AT sturmertil reducingbiasamplificationinthepresenceofunmeasuredconfoundingthroughoutofsampleestimationstrategiesforthediseaseriskscore |