Prognostic score–based balance measures can be a useful diagnostic for propensity score methods in comparative effectiveness research

Stuart, Elizabeth A.; Lee, Brian K.; Leacy, Finbarr P.

doi:10.1016/j.jclinepi.2013.01.013

Cited by 509 publications

(401 citation statements)

References 28 publications

Supporting

Mentioning

378

Contrasting

Order By: Relevance

“…Errors or missing data are inevitable; however, they would tend to bias toward the null hypothesis. Despite the use of a rigorous statistical approach, 25 there are no ways to adjust for unmeasured variables. Also, individuals receiving immunosuppression are followed closely, whereas those doing well may have been excluded from the study or prematurely lost to follow-up.…”

Section: Discussionmentioning

confidence: 99%

Corticosteroids in IgA Nephropathy

Tesař

Troyanov

Bellur

et al. 2015

Journal of the American Society of Nephrology

198

161

View full text Add to dashboard Cite

Current guidelines suggest treatment with corticosteroids (CS) in IgA nephropathy (IgAN) when proteinuria is persistently $1 g/d despite 3-6 months of supportive care and when eGFR is .50 ml/min per 1.73 m 2 . Whether the benefits of this treatment extend to patients with an eGFR#50 ml/min per 1.73 m 2 , other levels of proteinuria, or different renal pathologic lesions remains unknown. We retrospectively studied 1147 patients with IgAN from the European Validation Study of the Oxford Classification of IgAN (VALIGA) cohort classified according to the Oxford-MEST classification and medication used, with details of duration but not dosing. Overall, 46% of patients received immunosuppression, of which 98% received CS. Treated individuals presented with greater clinical and pathologic risk factors of progression. They also received more antihypertensive medication, and a greater proportion received renin angiotensin system blockade (RASB) compared with individuals without immunosuppressive therapy. Immunosuppression was associated with a significant reduction in proteinuria, a slower rate of renal function decline, and greater renal survival. Using a propensity score, we matched 184 subjects who received CS and RASB to 184 patients with a similar risk profile of progression who received only RASB. Within this group, CS reduced proteinuria and the rate of renal function decline and increased renal survival. These benefits extended to those with an eGFR#50 ml/min per 1.73 m 2 , and the benefits increased proportionally with the level of proteinuria. Thus, CS reduced the risk of progression regardless of initial eGFR and in direct proportion to the extent of proteinuria in this cohort.

show abstract

Section: Discussionmentioning

confidence: 99%

Corticosteroids in IgA Nephropathy

Tesař

Troyanov

Bellur

et al. 2015

Journal of the American Society of Nephrology

198

161

View full text Add to dashboard Cite

show abstract

“…Although there is no gold standard for measuring balance, simulation and theoretical work has suggested that the SMD is a better measure of balance than others measures (e.g., t test statistic and Kolmogorov-Smirnov [KS]; Ho, Imai, King, & Stuart, 2007; Stuart et al, 2013). As such, we used the SMD's to assess balance.…”

Section: Methodsmentioning

confidence: 99%

Selection and Socialization Effects in Early Adolescent Alcohol Use: A Propensity Score Analysis

Scalco

Trucco

Coffman

et al. 2015

J Abnorm Child Psychol

View full text Add to dashboard Cite

The robust correlation between peer and adolescent alcohol use (AU) has been taken as evidence for both socialization and selection processes in the etiology of adolescent AU. Accumulating evidence from studies using a diverse range of methodological and statistical approaches suggests that both processes are involved. A major challenge in testing whether peer AU predicts an adolescent's drinking (socialization) or whether an adolescent's drinking predicts peer AU (selection) is the myriad of potentially confounding factors that might lead to an overestimation of socialization and selection effects. After creating AU transition groups based on peer and adolescent AU across two waves (N = 765; age = 10-15; 53% female), we test whether transitions into AU by adolescents and peers predict later peer and adolescent AU respectively, using (1) propensity score analysis to balance transition groups on 26 potential confounds, (2) a longitudinal design with three waves to establish temporal precedence, and (3) both adolescent (target) and peer self-report of peer AU to disentangle effects attributable to shared reporter bias. Both selection and socialization were supported using both peer self-report of AU and adolescent-report of peer AU. Although cross-sectional analyses suggested peer self-reported models were associated with smaller effects than perceived peer AU, longitudinal analyses suggest a similar sized effect across reporter of peer AU for both selection and socialization. The implications of these findings for the etiology and treatment of adolescent AU are discussed.

show abstract

“…First, the use of outcome information in one of the proposed MSE-based approaches should be carefully considered. Some recent research has strongly cautioned against the use of outcome information in the estimation of propensity score weights (Stuart et al, 2013; Rubin, 2004; Rosenbaum, 2010; Hansen, 2008). As described by Stuart et al (2013), propensity score methods tend to be conducted without use of the outcome variable in an effort to separate the design and analysis stages of a study and allow for use of a single set of propensity scores for multiple outcomes.…”

Section: Discussionmentioning

confidence: 99%

“…Some recent research has strongly cautioned against the use of outcome information in the estimation of propensity score weights (Stuart et al, 2013; Rubin, 2004; Rosenbaum, 2010; Hansen, 2008). As described by Stuart et al (2013), propensity score methods tend to be conducted without use of the outcome variable in an effort to separate the design and analysis stages of a study and allow for use of a single set of propensity scores for multiple outcomes. However, others have argued that without the use of outcome information, instrumental variables which are related to the treatment but not related to the outcome may be included in the propensity score model and result in decreased precision (Brookhart et al, 2006; Westreich et al, 2011).…”

Section: Discussionmentioning

confidence: 99%

Optimizing variance-bias trade-off in the TWANG package for estimation of propensity scores

Parast

McCaffrey

Burgette

et al. 2016

Health Serv Outcomes Res Method

View full text Add to dashboard Cite

While propensity score weighting has been shown to reduce bias in treatment effect estimation when selection bias is present, it has also been shown that such weighting can perform poorly if the estimated propensity score weights are highly variable. Various approaches have been proposed which can reduce the variability of the weights and the risk of poor performance, particularly those based on machine learning methods. In this study, we closely examine approaches to fine-tune one machine learning technique (generalized boosted models [GBM]) to select propensity scores that seek to optimize the variance-bias trade-off that is inherent in most propensity score analyses. Specifically, we propose and evaluate three approaches for selecting the optimal number of trees for the GBM in the package in R. Normally, the package in R iteratively selects the optimal number of trees as that which maximizes balance between the treatment groups being considered. Because the selected number of trees may lead to highly variable propensity score weights, we examine alternative ways to tune the number of trees used in the estimation of propensity score weights such that we sacrifice some balance on the pre-treatment covariates in exchange for less variable weights. We use simulation studies to illustrate these methods and to describe the potential advantages and disadvantages of each method. We apply these methods to two case studies: one examining the effect of dog ownership on the owner’s general health using data from a large, population-based survey in California, and a second investigating the relationship between abstinence and a long-term economic outcome among a sample of high-risk youth.

show abstract

Prognostic score–based balance measures can be a useful diagnostic for propensity score methods in comparative effectiveness research

Cited by 509 publications

References 28 publications

Corticosteroids in IgA Nephropathy

Corticosteroids in IgA Nephropathy

Selection and Socialization Effects in Early Adolescent Alcohol Use: A Propensity Score Analysis

Optimizing variance-bias trade-off in the TWANG package for estimation of propensity scores

Contact Info

Product

Resources

About