Transporting results in an observational epidemiology setting: purposes, methods, and applied example
In the medical domain, substantial effort has been invested in generating internally valid estimates in experimental as well as observational studies, but limited effort has been made in testing generalizability, or external validity. Testing the external validity of scientific findings is neverthel...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2024-02-01
|
| Series: | Frontiers in Epidemiology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fepid.2024.1335241/full |
| _version_ | 1797291679155224576 |
|---|---|
| author | Ghislaine Scelo Daniela Zugna Maja Popovic Katrine Strandberg-Larsen Lorenzo Richiardi |
| author_facet | Ghislaine Scelo Daniela Zugna Maja Popovic Katrine Strandberg-Larsen Lorenzo Richiardi |
| author_sort | Ghislaine Scelo |
| collection | DOAJ |
| description | In the medical domain, substantial effort has been invested in generating internally valid estimates in experimental as well as observational studies, but limited effort has been made in testing generalizability, or external validity. Testing the external validity of scientific findings is nevertheless crucial for the application of knowledge across populations. In particular, transporting estimates obtained from observational studies requires the combination of methods for causal inference and methods to transport the effect estimates in order to minimize biases inherent to observational studies and to account for differences between the study and target populations. In this paper, the conceptual framework and assumptions behind transporting results from a population-based study population to a target population is described in an observational setting. An applied example to life-course epidemiology, where internal validity was constructed for illustrative purposes, is shown by using the targeted maximum likelihood estimator. |
| first_indexed | 2024-03-07T19:40:14Z |
| format | Article |
| id | doaj.art-08b37a2a1cbb4d7981f93242c09a7f4c |
| institution | Directory Open Access Journal |
| issn | 2674-1199 |
| language | English |
| last_indexed | 2024-03-07T19:40:14Z |
| publishDate | 2024-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Epidemiology |
| spelling | doaj.art-08b37a2a1cbb4d7981f93242c09a7f4c2024-02-29T05:31:45ZengFrontiers Media S.A.Frontiers in Epidemiology2674-11992024-02-01410.3389/fepid.2024.13352411335241Transporting results in an observational epidemiology setting: purposes, methods, and applied exampleGhislaine Scelo0Daniela Zugna1Maja Popovic2Katrine Strandberg-Larsen3Lorenzo Richiardi4Department of Medical Sciences, University of Turin, CPO-Piemonte, Turin, ItalyDepartment of Medical Sciences, University of Turin, CPO-Piemonte, Turin, ItalyDepartment of Medical Sciences, University of Turin, CPO-Piemonte, Turin, ItalySection of Epidemiology, Department of Public Health, University of Copenhagen, Copenhagen, DenmarkDepartment of Medical Sciences, University of Turin, CPO-Piemonte, Turin, ItalyIn the medical domain, substantial effort has been invested in generating internally valid estimates in experimental as well as observational studies, but limited effort has been made in testing generalizability, or external validity. Testing the external validity of scientific findings is nevertheless crucial for the application of knowledge across populations. In particular, transporting estimates obtained from observational studies requires the combination of methods for causal inference and methods to transport the effect estimates in order to minimize biases inherent to observational studies and to account for differences between the study and target populations. In this paper, the conceptual framework and assumptions behind transporting results from a population-based study population to a target population is described in an observational setting. An applied example to life-course epidemiology, where internal validity was constructed for illustrative purposes, is shown by using the targeted maximum likelihood estimator.https://www.frontiersin.org/articles/10.3389/fepid.2024.1335241/fulltransportabilityexternal validityobservational researchbirth cohortsTMLE |
| spellingShingle | Ghislaine Scelo Daniela Zugna Maja Popovic Katrine Strandberg-Larsen Lorenzo Richiardi Transporting results in an observational epidemiology setting: purposes, methods, and applied example Frontiers in Epidemiology transportability external validity observational research birth cohorts TMLE |
| title | Transporting results in an observational epidemiology setting: purposes, methods, and applied example |
| title_full | Transporting results in an observational epidemiology setting: purposes, methods, and applied example |
| title_fullStr | Transporting results in an observational epidemiology setting: purposes, methods, and applied example |
| title_full_unstemmed | Transporting results in an observational epidemiology setting: purposes, methods, and applied example |
| title_short | Transporting results in an observational epidemiology setting: purposes, methods, and applied example |
| title_sort | transporting results in an observational epidemiology setting purposes methods and applied example |
| topic | transportability external validity observational research birth cohorts TMLE |
| url | https://www.frontiersin.org/articles/10.3389/fepid.2024.1335241/full |
| work_keys_str_mv | AT ghislainescelo transportingresultsinanobservationalepidemiologysettingpurposesmethodsandappliedexample AT danielazugna transportingresultsinanobservationalepidemiologysettingpurposesmethodsandappliedexample AT majapopovic transportingresultsinanobservationalepidemiologysettingpurposesmethodsandappliedexample AT katrinestrandberglarsen transportingresultsinanobservationalepidemiologysettingpurposesmethodsandappliedexample AT lorenzorichiardi transportingresultsinanobservationalepidemiologysettingpurposesmethodsandappliedexample |