| Type | Journal Article |
|---|---|
| Author | J. Wilkinson |
| Author | M. A. Mamas |
| Author | E. Kontopantelis |
| URL | https://www.jclinepi.com/article/S0895-4356(22)00230-X/fulltext?rss=yes |
| Volume | 0 |
| Issue | 0 |
| Publication | Journal of Clinical Epidemiology |
| ISSN | 0895-4356, 1878-5921 |
| Date | 2022-09-17 |
| Extra | Publisher: Elsevier PMID: 36126791 |
| Journal Abbr | Journal of Clinical Epidemiology |
| DOI | 10.1016/j.jclinepi.2022.09.009 |
| Accessed | 9/21/2022, 6:54:20 AM |
| Library Catalog | www.jclinepi.com |
| Language | English |
| Abstract | Objective In observational studies, researchers must select a method to control for confounding. Options include propensity score methods and regression. It remains unclear how dataset characteristics (size, overlap in propensity scores, exposure prevalence) influence the relative performance of the methods. Study Design A simulation study to evaluate the role of dataset characteristics on the performance of propensity score methods, compared to logistic regression, for estimating a marginal odds ratio was conducted. Dataset size, overlap in propensity scores, and exposure prevalence were varied. Results Regression showed poor coverage for small sample sizes, but with large sample sizes was relatively robust to imbalance in propensity scores and low exposure prevalence. Propensity score methods displayed suboptimal coverage as overlap in propensity scores decreased, which was exacerbated at larger sample sizes. Power of matching methods was particularly affected by lack of overlap, low exposure prevalence, and small sample size. The advantage of regression for large data size was reduced in sensitivity analysis with a complementary log-log outcome generation mechanism and unmeasured confounding, with superior bias and error but inferior coverage to matching methods. Conclusion Dataset characteristics influence performance of methods for confounder adjustment. In many scenarios, regression may be the preferable option. |
| Short Title | Dataset characteristics affected the performance of propensity score methods for controlling confounding |
| Date Added | 9/21/2022, 6:54:20 AM |
| Modified | 9/21/2022, 6:55:35 AM |
| Type | Journal Article |
|---|---|
| Author | Fredrik Sävje |
| URL | https://doi.org/10.1093/biomet/asab035 |
| Issue | asab035 |
| Publication | Biometrika |
| ISSN | 0006-3444 |
| Date | June 22, 2021 |
| Journal Abbr | Biometrika |
| DOI | 10.1093/biomet/asab035 |
| Accessed | 6/23/2021, 4:44:03 PM |
| Library Catalog | Silverchair |
| Abstract | The paper shows that matching without replacement on propensity scores produces estimators that generally are inconsistent for the average treatment effect of the treated. To achieve consistency, practitioners must either assume that no units exist with propensity scores greater than one-half or assume that there is no confounding among such units. The result is not driven by the use of propensity scores, and similar artifacts arise when matching on other scores as long as it is without replacement. |
| Date Added | 6/23/2021, 4:44:03 PM |
| Modified | 6/23/2021, 4:44:28 PM |
| Type | Journal Article |
|---|---|
| Author | Markus C. Elze |
| Author | John Gregson |
| Author | Usman Baber |
| Author | Elizabeth Williamson |
| Author | Samantha Sartori |
| Author | Roxana Mehran |
| Author | Melissa Nichols |
| Author | Gregg W. Stone |
| Author | Stuart J. Pocock |
| URL | https://www.sciencedirect.com/science/article/pii/S073510971637036X |
| Volume | 69 |
| Issue | 3 |
| Pages | 345-357 |
| Publication | Journal of the American College of Cardiology |
| ISSN | 0735-1097 |
| Date | January 24, 2017 |
| Journal Abbr | Journal of the American College of Cardiology |
| DOI | 10.1016/j.jacc.2016.10.060 |
| Accessed | 3/1/2021, 4:08:01 PM |
| Library Catalog | ScienceDirect |
| Language | en |
| Abstract | Propensity scores (PS) are an increasingly popular method to adjust for confounding in observational studies. Propensity score methods have theoretical advantages over conventional covariate adjustment, but their relative performance in real-word scenarios is poorly characterized. We used datasets from 4 large-scale cardiovascular observational studies (PROMETHEUS, ADAPT-DES [the Assessment of Dual AntiPlatelet Therapy with Drug-Eluting Stents], THIN [The Health Improvement Network], and CHARM [Candesartan in Heart Failure-Assessment of Reduction in Mortality and Morbidity]) to compare the performance of conventional covariate adjustment with 4 common PS methods: matching, stratification, inverse probability weighting, and use of PS as a covariate. We found that stratification performed poorly with few outcome events, and inverse probability weighting gave imprecise estimates of treatment effect and undue influence to a small number of observations when substantial confounding was present. Covariate adjustment and matching performed well in all of our examples, although matching tended to give less precise estimates in some cases. PS methods are not necessarily superior to conventional covariate adjustment, and care should be taken to select the most suitable method. |
| Short Title | Comparison of Propensity Score Methods and Covariate Adjustment |
| Date Added | 3/1/2021, 4:08:01 PM |
| Modified | 3/1/2021, 4:20:40 PM |
| Type | Journal Article |
|---|---|
| Author | Stephen Senn |
| Author | Erika Graf |
| Author | Angelika Caputo |
| Volume | 26 |
| Pages | 5529-5544 |
| Publication | Stat Med |
| Date | 2007 |
| Extra | Citation Key: sen07str tex.citeulike-article-id= 13265650 tex.posted-at= 2014-07-14 14:10:01 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 1/19/2021, 8:03:32 AM |
propensity adjustment puts all emphasis on bias at the cost of variance
| Type | Journal Article |
|---|---|
| Author | Helen A. Blake |
| Author | Clémence Leyrat |
| Author | Kathryn E. Mansfield |
| Author | Shaun Seaman |
| Author | Laurie A. Tomlinson |
| Author | James Carpenter |
| Author | Elizabeth J. Williamson |
| URL | https://onlinelibrary.wiley.com/doi/abs/10.1002/sim.8503 |
| Rights | © 2020 John Wiley & Sons, Ltd. |
| Volume | n/a |
| Issue | n/a |
| Publication | Statistics in Medicine |
| ISSN | 1097-0258 |
| Date | 2020 |
| Extra | _eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1002/sim.8503 |
| DOI | 10.1002/sim.8503 |
| Accessed | 2/28/2020, 4:02:36 PM |
| Library Catalog | Wiley Online Library |
| Language | en |
| Abstract | Electronic health records are a valuable data source for investigating health-related questions, and propensity score analysis has become an increasingly popular approach to address confounding bias in such investigations. However, because electronic health records are typically routinely recorded as part of standard clinical care, there are often missing values, particularly for potential confounders. In our motivating study—using electronic health records to investigate the effect of renin-angiotensin system blockers on the risk of acute kidney injury—two key confounders, ethnicity and chronic kidney disease stage, have 59% and 53% missing data, respectively. The missingness pattern approach (MPA), a variant of the missing indicator approach, has been proposed as a method for handling partially observed confounders in propensity score analysis. In the MPA, propensity scores are estimated separately for each missingness pattern present in the data. Although the assumptions underlying the validity of the MPA are stated in the literature, it can be difficult in practice to assess their plausibility. In this article, we explore the MPA's underlying assumptions by using causal diagrams to assess their plausibility in a range of simple scenarios, drawing general conclusions about situations in which they are likely to be violated. We present a framework providing practical guidance for assessing whether the MPA's assumptions are plausible in a particular setting and thus deciding when the MPA is appropriate. We apply our framework to our motivating study, showing that the MPA's underlying assumptions appear reasonable, and we demonstrate the application of MPA to this study. |
| Short Title | Propensity scores using missingness pattern information |
| Date Added | 2/28/2020, 4:02:36 PM |
| Modified | 2/28/2020, 4:03:14 PM |
| Type | Journal Article |
|---|---|
| Author | Corwin Matthew Zigler |
| URL | https://doi.org/10.1080/00031305.2015.1111260 |
| Volume | 70 |
| Issue | 1 |
| Pages | 47-54 |
| Publication | The American Statistician |
| ISSN | 0003-1305 |
| Date | January 2, 2016 |
| Extra | PMID: 27482121 |
| DOI | 10.1080/00031305.2015.1111260 |
| Accessed | 11/25/2019, 7:36:13 AM |
| Library Catalog | Taylor and Francis+NEJM |
| Abstract | Although propensity scores have been central to the estimation of causal effects for over 30 years, only recently has the statistical literature begun to consider in detail methods for Bayesian estimation of propensity scores and causal effects. Underlying this recent body of literature on Bayesian propensity score estimation is an implicit discordance between the goal of the propensity score and the use of Bayes’ theorem. The propensity score condenses multivariate covariate information into a scalar to allow estimation of causal effects without specifying a model for how each covariate relates to the outcome. Avoiding specification of a detailed model for the outcome response surface is valuable for robust estimation of causal effects, but this strategy is at odds with the use of Bayes’ theorem, which presupposes a full probability model for the observed data that adheres to the likelihood principle. The goal of this article is to explicate this fundamental feature of Bayesian estimation of causal effects with propensity scores to provide context for the existing literature and for future work on this important topic.[Received June 2014. Revised September 2015.] |
| Date Added | 11/25/2019, 7:36:13 AM |
| Modified | 11/25/2019, 7:37:12 AM |
| Type | Journal Article |
|---|---|
| Author | Baiju R. Shah |
| Author | Andreas Laupacis |
| Author | Janet E. Hux |
| Author | Peter C. Austin |
| Volume | 58 |
| Pages | 550-559 |
| Publication | J Clin Epi |
| Date | 2005 |
| Extra | Citation Key: sha05pro tex.citeulike-article-id= 13265415 tex.posted-at= 2014-07-14 14:09:56 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
"Propensity scores gave slightly weaker associations; however many of the reviewed studies did not implement propensity scores well.";e.g. many papers tried to be parsimoneous in the propensity model
| Type | Journal Article |
|---|---|
| Author | David R. Judkins |
| Author | David Morganstein |
| Author | Paul Zador |
| Author | Andrea Piesse |
| Author | Brandon Barrett |
| Author | Pushpal Mukhopadhyay |
| Volume | 26 |
| Pages | 1022-1033 |
| Publication | Stat Med |
| Date | 2007 |
| Extra | Citation Key: jud07var tex.citeulike-article-id= 13265561 tex.posted-at= 2014-07-14 14:09:59 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
"articles dating back ... have pointed to situations where variance reduction can be achieved by making the models richer than strictly required to achieve bias elimination. In practice, this suggests a double benefit to overfitting in that both the risk of bias and variance are made smaller. We show that this double benefit is available only when the number of potential covariates is small relative to the sample size and when the covariates are related to outcomes.";ordinal treatment intensity modeled with ordinal logistic model;overfitting of the propensity model is beneficial if the extra variables are correlated with the outcome;if there are "only 10 cases per available covariate, dramatic overfitting can easily occur that will rob the analysis of all power to detect effects";polytomous model when PO violated;raking requires large sample sizes and categorical covariates
| Type | Journal Article |
|---|---|
| Author | Ben B. Hansen |
| Volume | 95 |
| Pages | 481-488 |
| Publication | Biometrika |
| Date | 2008 |
| Extra | Citation Key: han08pro tex.citeulike-article-id= 13265679 tex.posted-at= 2014-07-14 14:10:01 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
prognostic score (Miettinen multivariate confounder score) is especially prone to overfitting;problems with same sample estimation
| Type | Journal Article |
|---|---|
| Author | Lawrence C. McCandless |
| Author | Paul Gustafson |
| Author | Peter C. Austin |
| Volume | 28 |
| Pages | 94-112 |
| Publication | Stat Med |
| Date | 2009 |
| Extra | Citation Key: mcc09bay tex.citeulike-article-id= 13265723 tex.posted-at= 2014-07-14 14:10:02 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
using Bayesian credible intervals to adjust for uncertainty in estimation of propensity score;relied heavily on Rubin 5-category propensity adjustment
| Type | Journal Article |
|---|---|
| Author | James Robins |
| Volume | 8 |
| Pages | 679-701 |
| Publication | Stat Med |
| Date | 1989 |
| Extra | Citation Key: rob89con tex.citeulike-article-id= 13264738 tex.posted-at= 2014-07-14 14:09:41 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Paul R. Rosenbaum |
| Author | Donald B. Rubin |
| Volume | 45 |
| Pages | 212-218 |
| Publication | J Roy Stat Soc B |
| Date | 1983 |
| Extra | Citation Key: ros83ass tex.citeulike-article-id= 13264750 tex.posted-at= 2014-07-14 14:09:41 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | P. R. Rosenbaum |
| Author | D. Rubin |
| Volume | 70 |
| Pages | 41-55 |
| Publication | Biometrika |
| Date | 1983 |
| Extra | Citation Key: ros83cen tex.citeulike-article-id= 13264751 tex.posted-at= 2014-07-14 14:09:41 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Paul R. Rosenbaum |
| Author | Donald B. Rubin |
| Volume | 39 |
| Pages | 33-38 |
| Publication | Am Statistician |
| Date | 1985 |
| Extra | Citation Key: ros85con tex.citeulike-article-id= 13264754 tex.posted-at= 2014-07-14 14:09:41 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | James M. Robins |
| Author | Steven D. Mark |
| Author | Whitney K. Newey |
| Volume | 48 |
| Pages | 479-495 |
| Publication | Biometrics |
| Date | 1992 |
| Extra | Citation Key: rob92esti tex.citeulike-article-id= 13264741 tex.posted-at= 2014-07-14 14:09:41 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Book |
|---|---|
| Author | Paul R. Rosenbaum |
| Place | New York |
| Publisher | Springer-Verlag |
| Date | 1995 |
| Extra | Citation Key: ros95obs tex.citeulike-article-id= 13264762 tex.posted-at= 2014-07-14 14:09:41 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Donald B. Rubin |
| Author | Neal Thomas |
| Volume | 52 |
| Pages | 249-264 |
| Publication | Biometrics |
| Date | 1996 |
| Extra | Citation Key: rub96mat tex.citeulike-article-id= 13264778 tex.posted-at= 2014-07-14 14:09:42 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Donald B. Rubin |
| Volume | 127 |
| Pages | 757-763 |
| Publication | Ann Int Med |
| Date | 1997 |
| Extra | Citation Key: rub97est tex.citeulike-article-id= 13264779 tex.posted-at= 2014-07-14 14:09:42 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | J. Peter Weiss |
| Author | Carol Gruver |
| Author | Sanjiv Kaul |
| Author | Frank E. Harrell |
| Author | Jiri Sklenar |
| Author | John M. Dent |
| Volume | 13 |
| Pages | 124-130 |
| Publication | J Am Soc Echocard |
| Date | 2000 |
| Extra | Citation Key: wei00ord tex.citeulike-article-id= 13265113 tex.posted-at= 2014-07-14 14:09:49 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Daniel Westreich |
| Author | Justin Lessler |
| Author | Michele J. Funk |
| Volume | 63 |
| Pages | 826-833 |
| Publication | J Clin Epi |
| Date | 2010 |
| Extra | Citation Key: wes10pro tex.citeulike-article-id= 13265826 tex.posted-at= 2014-07-14 14:10:05 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
use of more advanced predictive methods in developing propensity scores
| Type | Book |
|---|---|
| Author | Andrew Gelman |
| Author | Jennifer Hill |
| URL | http://www.amazon.com/exec/obidos/redirect?tag=citeulike07-20&path=ASIN/052168689X |
| Edition | 1 |
| Publisher | Cambridge University Press |
| ISBN | 0-521-68689-X |
| Date | 2006-12 |
| Extra | Citation Key: gel06dat tex.citeulike-article-id= 1334704 tex.citeulike-linkout-0= http://www.amazon.ca/exec/obidos/redirect?tag=citeulike09-20&path=ASIN/052168689X tex.citeulike-linkout-1= http://www.amazon.de/exec/obidos/redirect?tag=citeulike01-21&path=ASIN/052168689X tex.citeulike-linkout-2= http://www.amazon.fr/exec/obidos/redirect?tag=citeulike06-21&path=ASIN/052168689X tex.citeulike-linkout-3= http://www.amazon.jp/exec/obidos/ASIN/052168689X tex.citeulike-linkout-4= http://www.amazon.co.uk/exec/obidos/ASIN/052168689X/citeulike00-21 tex.citeulike-linkout-5= http://www.amazon.com/exec/obidos/redirect?tag=citeulike07-20&path=ASIN/052168689X tex.citeulike-linkout-6= http://www.worldcat.org/isbn/052168689X tex.citeulike-linkout-7= http://books.google.com/books?vid=ISBN052168689X tex.citeulike-linkout-8= http://www.amazon.com/gp/search?keywords=052168689X&index=books&linkCode=qs tex.citeulike-linkout-9= http://www.librarything.com/isbn/052168689X tex.day= 18 tex.howpublished= Paperback tex.posted-at= 2016-12-04 21:00:49 tex.priority= 2 |
| Abstract | Data Analysis Using Regression and Multilevel/Hierarchical Models is a comprehensive manual for the applied researcher who wants to perform data analysis using linear and nonlinear regression and multilevel models. The book introduces a wide variety of models, whilst at the same time instructing the reader in how to fit these models using available software packages. The book illustrates the concepts by working through scores of real data examples that have arisen from the authors’ own applied research, with programming codes provided for each one. Topics covered include causal inference, including regression, poststratification, matching, regression discontinuity, and instrumental variables, as well as multilevel logistic regression and missing-data imputation. Practical tips regarding building, fitting, and understanding are provided throughout. |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Jessica M. Franklin |
| Author | Wesley Eddings |
| Author | Peter C. Austin |
| Author | Elizabeth A. Stuart |
| Author | Sebastian Schneeweiss |
| URL | http://dx.doi.org/10.1002/sim.7250 |
| Pages | n/a |
| Publication | Stat Med |
| Date | 2017-01 |
| Extra | Citation Key: fra17com tex.citeulike-article-id= 14281729 tex.citeulike-attachment-1= fra17com.pdf; /pdf/user/harrelfe/article/14281729/1102672/fra17com.pdf; 1f5fe6204e1398690f6b3f39020f8561d32247ad tex.citeulike-linkout-0= http://dx.doi.org/10.1002/sim.7250 tex.day= 1 tex.posted-at= 2017-02-17 13:30:43 tex.priority= 2 |
| DOI | 10.1002/sim.7250 |
| Abstract | Nonrandomized studies of treatments from electronic healthcare databases are critical for producing the evidence necessary to making informed treatment decisions, but often rely on comparing rates of events observed in a small number of patients. In addition, studies constructed from electronic healthcare databases, for example, administrative claims data, often adjust for many, possibly hundreds, of potential confounders. Despite the importance of maximizing efficiency when there are many confounders and few observed outcome events, there has been relatively little research on the relative performance of different propensity score methods in this context. In this paper, we compare a wide variety of propensity-based estimators of the marginal relative risk. In contrast to prior research that has focused on specific statistical methods in isolation of other analytic choices, we instead consider a method to be defined by the complete multistep process from propensity score modeling to final treatment effect estimation. Propensity score model estimation methods considered include ordinary logistic regression, Bayesian logistic regression, lasso, and boosted regression trees. Methods for utilizing the propensity score include pair matching, full matching, decile strata, fine strata, regression adjustment using one or two nonlinear splines, inverse propensity weighting, and matching weights. We evaluate methods via a 'plasmode' simulation study, which creates simulated datasets on the basis of a real cohort study of two treatments constructed from administrative claims data. Our results suggest that regression adjustment and matching weights, regardless of the propensity score model estimation method, provide lower bias and mean squared error in the context of rare binary outcomes. Copyright 2017 John Wiley & Sons, Ltd. |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
Included regression adjustment for spline of propensity in the comparison, which came out as co-winner with weighted matching.
| Type | Journal Article |
|---|---|
| Author | Richard K. Crump |
| Author | V. Joseph Hotz |
| Author | Guido W. Imbens |
| Author | Oscar A. Mitnik |
| URL | http://dx.doi.org/10.1093/biomet/asn055 |
| Volume | 96 |
| Issue | 1 |
| Pages | asn055-199 |
| Publication | Biometrika |
| ISSN | 1464-3510 |
| Date | 2009-01 |
| Extra | Citation Key: cru09dea tex.citeulike-article-id= 4172464 tex.citeulike-attachment-1= cru09dea.pdf; /pdf/user/harrelfe/article/4172464/1001962/cru09dea.pdf; fa7879fde4385c8bca6f88002517ffcaac5e40d9 tex.citeulike-linkout-0= http://dx.doi.org/10.1093/biomet/asn055 tex.citeulike-linkout-1= http://biomet.oxfordjournals.org/content/96/1/187.abstract tex.citeulike-linkout-2= http://biomet.oxfordjournals.org/content/96/1/187.full.pdf tex.citeulike-linkout-3= http://www.ingentaconnect.com/content/oup/biomet/2009/00000096/00000001/art00014 tex.day= 24 tex.posted-at= 2015-01-22 13:53:54 tex.priority= 2 tex.publisher= Oxford University Press |
| DOI | 10.1093/biomet/asn055 |
| Abstract | Estimation of average treatment effects under unconfounded or ignorable treatment assignment is often hampered by lack of overlap in the covariate distributions between treatment groups. This lack of overlap can lead to imprecise estimates, and can make commonly used estimators sensitive to the choice of specification. In such cases researchers have often used ad hoc methods for trimming the sample. We develop a systematic approach to addressing lack of overlap. We characterize optimal subsamples for which the average treatment effect can be estimated most precisely. Under some conditions, the optimal selection rules depend solely on the propensity score. For a wide range of distributions, a good approximation to the optimal rule is provided by the simple rule of thumb to discard all units with estimated propensity scores outside the range [0.1,0.9]. |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Jessica M. Franklin |
| Author | Wesley Eddings |
| Author | Robert J. Glynn |
| Author | Sebastian Schneeweiss |
| URL | http://dx.doi.org/10.1093/aje/kwv108 |
| Volume | 182 |
| Issue | 7 |
| Pages | 651-659 |
| Publication | Am J Epi |
| ISSN | 1476-6256 |
| Date | 2015-10 |
| Extra | Citation Key: fra15reg tex.citeulike-article-id= 13898359 tex.citeulike-attachment-1= fra15reg.pdf; /pdf/user/harrelfe/article/13898359/1048898/fra15reg.pdf; 39307cda8e267a66c86bee376049380256c9fc8c tex.citeulike-linkout-0= http://dx.doi.org/10.1093/aje/kwv108 tex.citeulike-linkout-1= http://aje.oxfordjournals.org/content/early/2015/08/01/aje.kwv108.abstract tex.citeulike-linkout-2= http://aje.oxfordjournals.org/content/early/2015/08/01/aje.kwv108.full.pdf tex.citeulike-linkout-3= http://view.ncbi.nlm.nih.gov/pubmed/26233956 tex.citeulike-linkout-4= http://www.hubmed.org/display.cgi?uids=26233956 tex.day= 1 tex.pmid= 26233956 tex.posted-at= 2016-01-05 23:01:42 tex.priority= 0 tex.publisher= Oxford University Press |
| DOI | 10.1093/aje/kwv108 |
| Abstract | Selection and measurement of confounders is critical for successful adjustment in nonrandomized studies. Although the principles behind confounder selection are now well established, variable selection for confounder adjustment remains a difficult problem in practice, particularly in secondary analyses of databases. We present a simulation study that compares the high-dimensional propensity score algorithm for variable selection with approaches that utilize direct adjustment for all potential confounders via regularized regression, including ridge regression and lasso regression. Simulations were based on 2 previously published pharmacoepidemiologic cohorts and used the plasmode simulation framework to create realistic simulated data sets with thousands of potential confounders. Performance of methods was evaluated with respect to bias and mean squared error of the estimated effects of a binary treatment. Simulation scenarios varied the true underlying outcome model, treatment effect, prevalence of exposure and outcome, and presence of unmeasured confounding. Across scenarios, high-dimensional propensity score approaches generally performed better than regularized regression approaches. However, including the variables selected by lasso regression in a regular propensity score model also performed well and may provide a promising alternative variable selection method. The Author 2015. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
plasmode simulation; references Bross method of simultaneously scoring association with exposure, association with outcome, and prevalence
| Type | Journal Article |
|---|---|
| Author | Qingxia Chen |
| Author | Hui Nian |
| Author | Yuwei Zhu |
| Author | H. Keipp Talbot |
| Author | Marie R. Griffin |
| Author | Frank E. Harrell |
| URL | http://dx.doi.org/10.1002/sim.7021 |
| Volume | 35 |
| Issue | 25 |
| Pages | 4546-4558 |
| Publication | Stat Med |
| ISSN | 02776715 |
| Date | 2016-11 |
| Extra | Citation Key: che16too tex.citeulike-article-id= 14218257 tex.citeulike-attachment-1= che16too.pdf; /pdf/user/harrelfe/article/14218257/1093467/che16too.pdf; 3227b989b9dc8141aeff04faa3e49f67159e8358 tex.citeulike-linkout-0= http://dx.doi.org/10.1002/sim.7021 tex.day= 10 tex.posted-at= 2016-12-01 14:22:08 tex.priority= 2 |
| DOI | 10.1002/sim.7021 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Tibor Schuster |
| Author | Wilfrid K. Lowe |
| Author | Robert W. Platt |
| URL | http://dx.doi.org/10.1016/j.jclinepi.2016.05.017 |
| Publication | J Clin Epi |
| ISSN | 08954356 |
| Date | 2016-09 |
| Extra | Citation Key: sch17pro tex.citeulike-article-id= 14220515 tex.citeulike-linkout-0= http://dx.doi.org/10.1016/j.jclinepi.2016.05.017 tex.posted-at= 2016-12-06 02:38:48 tex.priority= 2 |
| DOI | 10.1016/j.jclinepi.2016.05.017 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Erinn M. Hade |
| Author | Bo Lu |
| URL | http://dx.doi.org/10.1002/sim.5884 |
| Volume | 33 |
| Issue | 1 |
| Pages | 74-87 |
| Publication | Stat Med |
| ISSN | 1097-0258 |
| Date | 2014-01 |
| Extra | Citation Key: had14bia tex.citeulike-article-id= 12445750 tex.citeulike-attachment-1= had14bia.pdf; /pdf/user/harrelfe/article/12445750/996108/had14bia.pdf; e432263d6788007ac57fbc904f87dbcc4ccad464 tex.citeulike-linkout-0= http://dx.doi.org/10.1002/sim.5884 tex.citeulike-linkout-1= http://view.ncbi.nlm.nih.gov/pubmed/23787715 tex.citeulike-linkout-2= http://www.hubmed.org/display.cgi?uids=23787715 tex.day= 15 tex.pmid= 23787715 tex.posted-at= 2014-11-29 16:49:33 tex.priority= 2 |
| DOI | 10.1002/sim.5884 |
| Abstract | The use of propensity score methods to adjust for selection bias in observational studies has become increasingly popular in public health and medical research. A substantial portion of studies using propensity score adjustment treat the propensity score as a conventional regression predictor. Through a Monte Carlo simulation study, Austin and colleagues. investigated the bias associated with treatment effect estimation when the propensity score is used as a covariate in nonlinear regression models, such as logistic regression and Cox proportional hazards models. We show that the bias exists even in a linear regression model when the estimated propensity score is used and derive the explicit form of the bias. We also conduct an extensive simulation study to compare the performance of such covariate adjustment with propensity score stratification, propensity score matching, inverse probability of treatment weighted method, and nonparametric functional estimation using splines. The simulation scenarios are designed to reflect real data analysis practice. Instead of specifying a known parametric propensity score model, we generate the data by considering various degrees of overlap of the covariate distributions between treated and control groups. Propensity score matching excels when the treated group is contained within a larger control pool, while the model-based adjustment may have an edge when treated and control groups do not have too much overlap. Overall, adjusting for the propensity score through stratification or matching followed by regression or using splines, appears to be a good practical strategy. |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Peter C. Austin |
| Volume | 61 |
| Pages | 537-545 |
| Publication | J Clin Epi |
| Date | 2008 |
| Extra | Citation Key: aus08per tex.citeulike-article-id= 13265663 tex.posted-at= 2014-07-14 14:10:01 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
MSE and bias of relative risk;quintile stratification vs. matching;covariate adjustment;"propensity-score matching resulted in estimates with less bias than did stratification on the quintiles of the propensity score, but stratification on the quintiles of the propensity score resulted in estimates with greater precision."
| Type | Journal Article |
|---|---|
| Author | Mi-Ja Woo |
| Author | Jerome P. Reiter |
| Author | Alan F. Karr |
| Volume | 27 |
| Pages | 3805-3816 |
| Publication | Stat Med |
| Date | 2008 |
| Extra | Citation Key: woo08est tex.citeulike-article-id= 13265701 tex.posted-at= 2014-07-14 14:10:02 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Donald B. Rubin |
| Volume | 2 |
| Issue | 3 |
| Pages | 808-840 |
| Publication | Ann Appl Stat |
| Date | 2008 |
| Extra | Citation Key: rub08obj tex.citeulike-article-id= 13265708 tex.posted-at= 2014-07-14 14:10:02 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
making observational research more rigorous;observational studies as approximations of randomized experiments;design observational studies to approximate randomized trials;make sure key covariates are measured well
| Type | Journal Article |
|---|---|
| Author | Andrew C. Leon |
| Author | Donald Hedeker |
| Volume | 24 |
| Pages | 647-658 |
| Publication | Stat Med |
| Date | 2005 |
| Extra | Citation Key: leo05mix tex.citeulike-article-id= 13265408 tex.posted-at= 2014-07-14 14:09:56 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Jared K. Lunceford |
| Author | Marie Davidian |
| Volume | 23 |
| Pages | 2937-2960 |
| Publication | Stat Med |
| Date | 2004 |
| Extra | Citation Key: lun04str tex.citeulike-article-id= 13265389 tex.posted-at= 2014-07-14 14:09:55 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
"Theoretical and empirical results indicate that the popular version of stratification via estimated propensity scores based on within-stratum sample mean differences and a fixed number of strata can lead to biased inference due to residual confounding, and the effect of this bias becomes more serious with increasing sample size. Using more strata can increase the sample size at which the trade-off of bias and variability involved in efficiency takes place, but stratifying on quintiles seems to be the most popular approach in practice, even for substantial sample sizes. Thus, as the trade-off point will be unknown for any specific problem, this approach should be used with caution. An interesting avenue for future research would be to establish guidelines for choosing the number of strata based on theoretical analysis of the rate at which the number of strata should increase with sample size to eliminate bias. A modification of stratification based instead on within-stratum regression estimates of treatment effect can eliminate this bias and achieve dramatic improvements in efficiency, but correct specification of the regression model is essential; otherwise, bias and degradation of performance can result. In this regard, this approach is similar to estimating causal effects via direct regression modelling but is less sensitive to mismodelling. Methods based on weighting are consistent and offer approximately unbiased inference for practical sample sizes. The semiparametric efficient estimator identified by the theory of Robins et al. [13], which incorporates regression modelling as a way to gain efficiency, also yields high precision. Although stratification based on regression and direct modelling can outperform this approach under some conditions, this estimator enjoys the unique 'double robustness' property in that it continues to lead to unbiased estimation of the average causal effect even if the regression models involved do not coincide with the true relationship, affording the analyst broad protection against misspecification not available with these other approaches. The results presented here support routine use of this estimator in practice."; nice derivations of inner workings of propensity score
| Type | Journal Article |
|---|---|
| Author | Olaf H. Klungel |
| Author | Edwin P. Martens |
| Author | Bruce M. Psaty |
| Author | Diederik E. Grobbee |
| Author | Sean D. Sullivan |
| Author | Bruno H. Stricker |
| Author | Hubert G. M. Leufkens |
| Author | A. de Boer |
| Volume | 57 |
| Pages | 1223-1231 |
| Publication | J Clin Epi |
| Date | 2004 |
| Extra | Citation Key: Klu04met tex.citeulike-article-id= 13265397 tex.posted-at= 2014-07-14 14:09:55 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Rusty Tchernis |
| Author | Marcela Horvitz-Lennon |
| Author | Sharon-Lise T. Normand |
| Volume | 24 |
| Pages | 2197-2212 |
| Publication | Stat Med |
| Date | 2005 |
| Extra | Citation Key: tch05use tex.citeulike-article-id= 13265437 tex.posted-at= 2014-07-14 14:09:56 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
multi-valued treatment propensity model using discrete choice model
| Type | Journal Article |
|---|---|
| Author | Peter C. Austin |
| Author | Muhammad M. Mamdani |
| Volume | 25 |
| Pages | 2084-2106 |
| Publication | Stat Med |
| Date | 2006 |
| Extra | Citation Key: aus06com tex.citeulike-article-id= 13265481 tex.posted-at= 2014-07-14 14:09:57 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
propensity adjustment using covariate adjustment;effect on hypothesis tests, relative and absolute effects;residual confounding remained after stratifying by quintiles of propensity;Q-Q plots were more sensitive for showing imbalances;non-collapsibility of odds ratios;matched pairs yielded estimates closer to the null because of this;allowing propensity effect on Y to be quadratic affected the result
| Type | Journal Article |
|---|---|
| Author | Peter C. Austin |
| Volume | 26 |
| Pages | 3078-3094 |
| Publication | Stat Med |
| Date | 2007 |
| Extra | Citation Key: aus07per tex.citeulike-article-id= 13265602 tex.posted-at= 2014-07-14 14:10:00 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Donald B. Rubin |
| Author | Neal Thomas |
| Volume | 95 |
| Pages | 573-585 |
| Publication | J Am Stat Assoc |
| Date | 2000 |
| Extra | Citation Key: rub00com tex.citeulike-article-id= 13265133 tex.posted-at= 2014-07-14 14:09:50 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | D'Agostino, Jr |
| Author | Donald B. Rubin |
| Volume | 95 |
| Pages | 749-759 |
| Publication | J Am Stat Assoc |
| Date | 2000 |
| Extra | Citation Key: dag00est tex.citeulike-article-id= 13265142 tex.posted-at= 2014-07-14 14:09:50 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
missing data in propensity score;adding extra categories for missing;using indicators of missingness in propensity score to balance on missing data patterns;general location model for likelihood-based analysis of partial data
| Type | Journal Article |
|---|---|
| Author | Sharon-Lise T. Normand |
| Author | Mary B. Landrum |
| Author | Edward Guadagnoli |
| Author | John Z. Ayanian |
| Author | Thomas J. Ryan |
| Author | Paul D. Cleary |
| Author | Barbara J. McNeil |
| Volume | 54 |
| Pages | 387-398 |
| Publication | J Clin Epi |
| Date | 2001 |
| Extra | Citation Key: nor01val tex.citeulike-article-id= 13265188 tex.posted-at= 2014-07-14 14:09:51 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Babette Brumback |
| Author | Sander Greenland |
| Author | Mary Redman |
| Author | Nancy Kiviat |
| Author | Paula Diehr |
| Volume | 59 |
| Pages | 274-285 |
| Publication | Biometrics |
| Date | 2003 |
| Extra | Citation Key: bru03int tex.citeulike-article-id= 13265334 tex.posted-at= 2014-07-14 14:09:54 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | G. Imbens |
| URL | http://konstanza.ingentaselect.com/vl=2884470/cl=22/nw=1/rpsv/ij/oup/00063444/v87n3/s17/p706 |
| Volume | 87 |
| Pages | 706-710 |
| Publication | Biometrika |
| Date | 2000 |
| Extra | Citation Key: imb00rol tex.citeulike-article-id= 13265382 tex.citeulike-linkout-0= http://konstanza.ingentaselect.com/vl=2884470/cl=22/nw=1/rpsv/ij/oup/00063444/v87n3/s17/p706 tex.posted-at= 2014-07-14 14:09:55 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | The SUPPORT Principal Investigators |
| URL | http://dx.doi.org/10.1001/jama.274.20.1591 |
| Volume | 274 |
| Pages | 1591-1598 |
| Publication | JAMA |
| Date | 1995 |
| Extra | Citation Key: sup95con tex.citeulike-article-id= 13264925 tex.citeulike-linkout-0= http://dx.doi.org/10.1001/jama.274.20.1591 tex.posted-at= 2014-07-14 14:09:45 tex.priority= 0 |
| DOI | 10.1001/jama.274.20.1591 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Mark McClellan |
| Author | Barbara J. McNeil |
| Author | Joseph P. Newhouse |
| Volume | 272 |
| Pages | 859-866 |
| Publication | JAMA |
| Date | 1994 |
| Extra | Citation Key: mcc94 tex.citeulike-article-id= 13264588 tex.posted-at= 2014-07-14 14:09:38 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Philip W. Lavori |
| Author | Martin B. Keller |
| Volume | 7 |
| Pages | 727-737 |
| Publication | Stat Med |
| Date | 1988 |
| Extra | Citation Key: lav88imp tex.citeulike-article-id= 13264468 tex.posted-at= 2014-07-14 14:09:36 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Philip W. Lavori |
| Author | Ree Dawson |
| Author | David Shera |
| Volume | 14 |
| Pages | 1913-1925 |
| Publication | Stat Med |
| Date | 1995 |
| Extra | Citation Key: lav95mul tex.citeulike-article-id= 13264470 tex.posted-at= 2014-07-14 14:09:36 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | D. Y. Lin |
| Author | B. M. Psaty |
| Author | R. A. Kronmal |
| Volume | 54 |
| Pages | 948-963 |
| Publication | Biometrics |
| Date | 1998 |
| Extra | Citation Key: lin98ass tex.citeulike-article-id= 13264529 tex.posted-at= 2014-07-14 14:09:37 tex.priority= 0 See important letter to the editor by Tyler VanderWeele BCS 64:645-649; 2008 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | D. B. Mark |
| Author | C. L. Nelson |
| Author | R. M. Califf |
| Author | F. E. Harrell |
| Author | K. L. Lee |
| Author | R. H. Jones |
| Author | D. F. Fortin |
| Author | R. S. Stack |
| Author | D. D. Glower |
| Author | L. R. Smith |
| Author | E. R. DeLong |
| Author | P. K. Smith |
| Author | J. G. Reves |
| Author | J. G. Jollis |
| Author | J. E. Tcheng |
| Author | L. H. Muhlbaier |
| Author | J. E. Lowe |
| Author | H. R. Phillips |
| Author | D. B. Pryor |
| Volume | 89 |
| Pages | 2015-2025 |
| Publication | Circ |
| Date | 1994 |
| Extra | Citation Key: mar94con tex.citeulike-article-id= 13264577 tex.posted-at= 2014-07-14 14:09:38 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Marshall M. Joffe |
| Author | Paul R. Rosenbaum |
| Volume | 150 |
| Pages | 327-333 |
| Publication | Am J Epi |
| Date | 1999 |
| Extra | Citation Key: jof99pro tex.citeulike-article-id= 13264363 tex.posted-at= 2014-07-14 14:09:33 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Joseph Gastwirth |
| Author | Abba Krieger |
| Author | Paul Rosenbaum |
| Volume | 48 |
| Pages | 313-315 |
| Publication | Am Statistician |
| Date | 1994 |
| Extra | Citation Key: gas94how tex.citeulike-article-id= 13264122 tex.posted-at= 2014-07-14 14:09:29 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
see follow up articles in am stat 51 no 2 may 1997
| Type | Journal Article |
|---|---|
| Author | William H. Crown |
| Author | Robert L. Obenchain |
| Author | Luella Englehart |
| Author | Tamra Lair |
| Author | Don P. Buesching |
| Author | Thomas Croghan |
| Volume | 17 |
| Pages | 1943-1958 |
| Publication | Stat Med |
| Date | 1998 |
| Extra | Citation Key: cro98app tex.citeulike-article-id= 13263952 tex.posted-at= 2014-07-14 14:09:26 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Ralph B. D'Agostino |
| Author | Heidy Kwan |
| Volume | 33 |
| Pages | AS95-AS105 |
| Publication | Med Care |
| Date | 1995 |
| Extra | Citation Key: dag95mea tex.citeulike-article-id= 13263965 tex.posted-at= 2014-07-14 14:09:26 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | D'Agostino, Jr |
| Volume | 17 |
| Pages | 2265-2281 |
| Publication | Stat Med |
| Date | 1998 |
| Extra | Citation Key: dag98pro tex.citeulike-article-id= 13263966 tex.posted-at= 2014-07-14 14:09:26 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Christiana Drake |
| Volume | 49 |
| Pages | 1231-1236 |
| Publication | Biometrics |
| Date | 1993 |
| Extra | Citation Key: dra93eff tex.citeulike-article-id= 13264017 tex.posted-at= 2014-07-14 14:09:27 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Alfred F. Connors |
| Author | Theodore Speroff |
| Author | Neal V. Dawson |
| Author | Charles Thomas |
| Author | Frank E. Harrell |
| Author | Douglas Wagner |
| Author | Norman Desbiens |
| Author | Lee Goldman |
| Author | Albert W. Wu |
| Author | Robert M. Califf |
| Author | William J. Fulkerson |
| Author | Humberto Vidaillet |
| Author | Steven Broste |
| Author | Paul Bellamy |
| Author | Joanne Lynn |
| Author | William A. Knaus |
| Author | The SUPPORT Investigators |
| URL | http://www.ncbi.nlm.nih.gov/pubmed/8782638 |
| Volume | 276 |
| Pages | 889-897 |
| Publication | JAMA |
| Date | 1996 |
| Extra | Citation Key: con96eff tex.citeulike-article-id= 13263914 tex.citeulike-linkout-0= http://www.ncbi.nlm.nih.gov/pubmed/8782638 tex.citeulike-linkout-1= http://dx.doi.org/ tex.posted-at= 2014-07-14 14:09:25 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | E. Francis Cook |
| Author | Lee Goldman |
| Volume | 127 |
| Pages | 626-639 |
| Publication | Am J Epi |
| Date | 1988 |
| Extra | Citation Key: coo88asy tex.citeulike-article-id= 13263916 tex.posted-at= 2014-07-14 14:09:25 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Andrew J. Copas |
| Author | Vern T. Farewell |
| Volume | 161 |
| Pages | 385-396 |
| Publication | J Roy Stat Soc A |
| Date | 1998 |
| Extra | Citation Key: cop98dea tex.citeulike-article-id= 13263929 tex.posted-at= 2014-07-14 14:09:25 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Donald B. Rubin |
| Volume | 26 |
| Pages | 20-36 |
| Publication | Stat Med |
| Date | 2007 |
| Extra | Citation Key: rub07des tex.citeulike-article-id= 13265546 tex.posted-at= 2014-07-14 14:09:59 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Philip W. Lavori |
| Author | Ree Dawson |
| Author | Timothy B. Mueller |
| Volume | 13 |
| Pages | 1089-1100 |
| Publication | Stat Med |
| Date | 1994 |
| Extra | Citation Key: lav94cau tex.citeulike-article-id= 13264469 tex.posted-at= 2014-07-14 14:09:36 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | Joshua D. Angrist |
| Author | Guido W. Imbens |
| Volume | 90 |
| Pages | 431-442 |
| Publication | J Am Stat Assoc |
| Date | 1995 |
| Extra | Citation Key: ang95two tex.citeulike-article-id= 13263703 tex.posted-at= 2014-07-14 14:09:21 tex.priority= 0 |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
| Type | Journal Article |
|---|---|
| Author | S. Vansteelandt |
| Author | R. M. Daniel |
| URL | http://dx.doi.org/10.1002/sim.6207 |
| Volume | 33 |
| Issue | 23 |
| Pages | 4053-4072 |
| Publication | Stat Med |
| ISSN | 02776715 |
| Date | 2014-10 |
| Extra | Citation Key: van14reg tex.citeulike-article-id= 13340175 tex.citeulike-attachment-1= van14reg.pdf; /pdf/user/harrelfe/article/13340175/982683/van14reg.pdf; 7b2d0837c91c0d47441b5588ee35f8a735dd2f7b tex.citeulike-linkout-0= http://dx.doi.org/10.1002/sim.6207 tex.day= 15 tex.posted-at= 2014-08-28 19:21:49 tex.priority= 4 |
| DOI | 10.1002/sim.6207 |
| Abstract | Propensity scores are widely adopted in observational research because they enable adjustment for high-dimensional confounders without requiring models for their association with the outcome of interest. The results of statistical analyses based on stratification, matching or inverse weighting by the propensity score are therefore less susceptible to model extrapolation than those based solely on outcome regression models. This is attractive because extrapolation in outcome regression models may be alarming, yet difficult to diagnose, when the exposed and unexposed individuals have very different covariate distributions. Standard regression adjustment for the propensity score forms an alternative to the aforementioned propensity score methods, but the benefits of this are less clear because it still involves modelling the outcome in addition to the propensity score. In this article, we develop novel insights into the properties of this adjustment method. We demonstrate that standard tests of the null hypothesis of no exposure effect (based on robust variance estimators), as well as particular standardised effects obtained from such adjusted regression models, are robust against misspecification of the outcome model when a propensity score model is correctly specified; they are thus not vulnerable to the aforementioned problem of extrapolation. We moreover propose efficient estimators for these standardised effects, which retain a useful causal interpretation even when the propensity score model is misspecified, provided the outcome regression model is correctly specified. |
| Date Added | 7/7/2018, 1:38:33 PM |
| Modified | 11/8/2019, 8:01:59 AM |
From first author in ASA Connections discussion:I share your concerns and also tend to favour regression adjustment for the propensity score for the following additional reasons:- it adjusts for residual confounding due to imperfect matching;- it makes more efficient use of the information in the data;- adjustment for both the propensity score and covariates can result in estimators with a double robustness property: they are valid if either the propensity score is correct or the covariates in the outcome model are correctly modelled.- matching is often preferred for simplicity, but I tend to believe that regression adjustment for the propensity score is much simpler because one loses the simplicity of matching whenever one wishes to adjust for residual confounding due to imperfect matching or wishes to obtain valid standard errors.
Note that, like matching methods, regression adjustment for the propensity score also prevents the dangers of model extrapolation in settings where the treated and untreated are very different in their observed covariate data, in the sense that- subjects with covariate values at which there are nearly only treated or non-treated individuals are down-weighted, as they contribute little or no information about treatment effect;- valid estimators of treatment effect can be obtained when the propensity score model is correct, even when the association between outcome and propensity score is misspecified. This follows from the aforementioned double robustness property.
| Type | Journal Article |
|---|---|
| Author | Emily Granger |
| Author | Jamie C. Sergeant |
| Author | Mark Lunt |
| URL | https://onlinelibrary.wiley.com/doi/abs/10.1002/sim.8355 |
| Rights | © 2019 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd. |
| Volume | 38 |
| Issue | 26 |
| Pages | 5120-5132 |
| Publication | Statistics in Medicine |
| ISSN | 1097-0258 |
| Date | 2019 |
| DOI | 10.1002/sim.8355 |
| Accessed | 10/21/2019, 8:31:09 AM |
| Library Catalog | Wiley Online Library |
| Language | en |
| Abstract | Overcoming bias due to confounding and missing data is challenging when analyzing observational data. Propensity scores are commonly used to account for the first problem and multiple imputation for the latter. Unfortunately, it is not known how best to proceed when both techniques are required. We investigate whether two different approaches to combining propensity scores and multiple imputation (Across and Within) lead to differences in the accuracy or precision of exposure effect estimates. Both approaches start by imputing missing values multiple times. Propensity scores are then estimated for each resulting dataset. Using the Across approach, the mean propensity score across imputations for each subject is used in a single subsequent analysis. Alternatively, the Within approach uses propensity scores individually to obtain exposure effect estimates in each imputation, which are combined to produce an overall estimate. These approaches were compared in a series of Monte Carlo simulations and applied to data from the British Society for Rheumatology Biologics Register. Results indicated that the Within approach produced unbiased estimates with appropriate confidence intervals, whereas the Across approach produced biased results and unrealistic confidence intervals. Researchers are encouraged to implement the Within approach when conducting propensity score analyses with incomplete data. |
| Date Added | 10/21/2019, 8:31:09 AM |
| Modified | 10/21/2019, 8:31:42 AM |
| Type | Journal Article |
|---|---|
| Author | Shomoita Alam |
| Author | Erica E. M. Moodie |
| Author | David A. Stephens |
| URL | https://onlinelibrary.wiley.com/doi/abs/10.1002/sim.8075 |
| Rights | © 2018 John Wiley & Sons, Ltd. |
| Volume | 0 |
| Issue | 0 |
| Publication | Statistics in Medicine |
| ISSN | 1097-0258 |
| Date | 2018 |
| DOI | 10.1002/sim.8075 |
| Accessed | 12/31/2018, 9:26:20 AM |
| Library Catalog | Wiley Online Library |
| Language | en |
| Abstract | In investigations of the effect of treatment on outcome, the propensity score is a tool to eliminate imbalance in the distribution of confounding variables between treatment groups. Recent work has suggested that Super Learner, an ensemble method, outperforms logistic regression in nonlinear settings; however, experience with real-data analyses tends to show overfitting of the propensity score model using this approach. We investigated a wide range of simulated settings of varying complexities including simulations based on real data to compare the performances of logistic regression, generalized boosted models, and Super Learner in providing balance and for estimating the average treatment effect via propensity score regression, propensity score matching, and inverse probability of treatment weighting. We found that Super Learner and logistic regression are comparable in terms of covariate balance, bias, and mean squared error (MSE); however, Super Learner is computationally very expensive thus leaving no clear advantage to the more complex approach. Propensity scores estimated by generalized boosted models were inferior to the other two estimation approaches. We also found that propensity score regression adjustment was superior to either matching or inverse weighting when the form of the dependence on the treatment on the outcome is correctly specified. |
| Short Title | Should a propensity score model be super? |
| Date Added | 12/31/2018, 9:26:20 AM |
| Modified | 12/31/2018, 9:27:50 AM |
| Type | Journal Article |
|---|---|
| Author | Fei Wan |
| URL | https://onlinelibrary.wiley.com/doi/abs/10.1002/sim.7976 |
| Rights | © 2018 John Wiley & Sons, Ltd. |
| Volume | 38 |
| Issue | 2 |
| Pages | 289-300 |
| Publication | Statistics in Medicine |
| ISSN | 1097-0258 |
| Date | 2019 |
| DOI | 10.1002/sim.7976 |
| Accessed | 12/16/2018, 8:07:57 PM |
| Library Catalog | Wiley Online Library |
| Language | en |
| Abstract | Propensity-score matching has been used widely in observational studies to balance confounders across treatment groups. However, whether matched-pairs analyses should be used as a primary approach is still in debate. We compared the statistical power and type 1 error rate for four commonly used methods of analyzing propensity-score–matched samples with continuous outcomes: (1) an unadjusted mixed-effects model, (2) an unadjusted generalized estimating method, (3) simple linear regression, and (4) multiple linear regression. Multiple linear regression had the highest statistical power among the four competing methods. We also found that the degree of intraclass correlation within matched pairs depends on the dissimilarity between the coefficient vectors of confounders in the outcome and treatment models. Multiple linear regression is superior to the unadjusted matched-pairs analyses for propensity-score–matched data. |
| Date Added | 12/16/2018, 8:07:57 PM |
| Modified | 12/16/2018, 8:08:46 PM |