Optimization-Based Stable Balancing Weights Versus Propensity Score Weighting for Samples With High Covariate Imbalance.

IF 2.4 4区医学 Q3 PHARMACOLOGY & PHARMACY

Pharmacoepidemiology and Drug Safety Pub Date : 2024-07-01 DOI:10.1002/pds.5864

Stuart R Wallace, Sachinkumar B Singh, Rebekah Blakney, Lexi Rene, Stephen S Johnston

{"title":"Optimization-Based Stable Balancing Weights Versus Propensity Score Weighting for Samples With High Covariate Imbalance.","authors":"Stuart R Wallace, Sachinkumar B Singh, Rebekah Blakney, Lexi Rene, Stephen S Johnston","doi":"10.1002/pds.5864","DOIUrl":null,"url":null,"abstract":"Purpose: To compare the performance (covariate balance, effective sample size [ESS]) of stable balancing weights (SBW) versus propensity score weighting (PSW). Two applied cases were used to compare performance: (Case 1) extreme imbalance in baseline covariates between groups and (Case 2) substantial discrepancy in sample size between groups.Methods: Using the Premier Healthcare Database, we selected patients who (Case 1) underwent a surgical procedure with one of two different bipolar forceps between January 2000 and June 2020, or (Case 2) a neurological procedure using one of two different nonabsorbable surgical sutures between January 2000 and March 2020. Average treatment effects on the treated (ATT) weights were generated based on selected covariates. SBW was implemented using two techniques: (1) \"grid search\" to find weights of minimum variance at the lowest target absolute standardized mean difference (SMD); (2) finding weights of minimum variance at prespecified SMD tolerance. PSW and SBW methods were compared on postweighting SMDs, the number of imbalanced covariates, and ESS of the ATT-weighted control group.Results: In both studies, improved covariate balance was achieved with both SBW techniques. All methods suffered from postweighting ESS that was lower than the unweighted control group's original sample size; however, SBW methods achieved higher ESS for the control groups. Sensitivity analyses using SBW to apply variable-specific SMD thresholds increased ESS, outperforming PSW.Conclusions: In this applied example, the optimization-based SBW method provided ample flexibility with respect to prespecification of covariate balance goals and resulted in better postweighting covariate balance and larger ESS as compared with PSW.","PeriodicalId":19782,"journal":{"name":"Pharmacoepidemiology and Drug Safety","volume":"33 7","pages":"e5864"},"PeriodicalIF":2.4000,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Pharmacoepidemiology and Drug Safety","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1002/pds.5864","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"PHARMACOLOGY & PHARMACY","Score":null,"Total":0}

引用次数: 0

Abstract

Purpose: To compare the performance (covariate balance, effective sample size [ESS]) of stable balancing weights (SBW) versus propensity score weighting (PSW). Two applied cases were used to compare performance: (Case 1) extreme imbalance in baseline covariates between groups and (Case 2) substantial discrepancy in sample size between groups.

Methods: Using the Premier Healthcare Database, we selected patients who (Case 1) underwent a surgical procedure with one of two different bipolar forceps between January 2000 and June 2020, or (Case 2) a neurological procedure using one of two different nonabsorbable surgical sutures between January 2000 and March 2020. Average treatment effects on the treated (ATT) weights were generated based on selected covariates. SBW was implemented using two techniques: (1) "grid search" to find weights of minimum variance at the lowest target absolute standardized mean difference (SMD); (2) finding weights of minimum variance at prespecified SMD tolerance. PSW and SBW methods were compared on postweighting SMDs, the number of imbalanced covariates, and ESS of the ATT-weighted control group.

Results: In both studies, improved covariate balance was achieved with both SBW techniques. All methods suffered from postweighting ESS that was lower than the unweighted control group's original sample size; however, SBW methods achieved higher ESS for the control groups. Sensitivity analyses using SBW to apply variable-specific SMD thresholds increased ESS, outperforming PSW.

Conclusions: In this applied example, the optimization-based SBW method provided ample flexibility with respect to prespecification of covariate balance goals and resulted in better postweighting covariate balance and larger ESS as compared with PSW.

查看原文本刊更多论文

基于优化的稳定平衡加权法与倾向得分加权法相比较，适用于共变量高度不平衡的样本。

目的：比较稳定平衡加权（SBW）与倾向得分加权（PSW）的性能（协变量平衡、有效样本量[ESS]）。比较性能时使用了两个应用案例：（案例 1）组间基线协变量极度不平衡；（案例 2）组间样本量存在巨大差异：我们使用 Premier Healthcare 数据库，选择了 2000 年 1 月至 2020 年 6 月期间（情况 1）使用两种不同的双极钳之一进行外科手术的患者，或 2000 年 1 月至 2020 年 3 月期间（情况 2）使用两种不同的非吸收性手术缝合线之一进行神经外科手术的患者。根据选定的协变量生成对治疗者的平均治疗效果（ATT）权重。SBW 采用两种技术：(1) "网格搜索"，在最低的目标绝对标准化均值差（SMD）上寻找方差最小的权重；(2) 在预先指定的 SMD 容差上寻找方差最小的权重。在加权后 SMD、不平衡协变量的数量以及 ATT 加权对照组的 ESS 方面，比较了 PSW 和 SBW 方法：结果：在这两项研究中，SBW 和 SBW 方法都改善了协变量的平衡。所有方法的加权后 ESS 都低于未加权对照组的原始样本量；但 SBW 方法的对照组 ESS 更高。使用 SBW 进行敏感性分析，应用特定变量的 SMD 临界值提高了 ESS，优于 PSW：在这个应用实例中，基于优化的 SBW 方法在预设协方差平衡目标方面提供了充分的灵活性，与 PSW 相比，SBW 方法可实现更好的加权后协方差平衡和更大的 ESS。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Pharmacoepidemiology and Drug Safety 医学-药学

CiteScore

4.80

自引率

7.70%

发文量

173

审稿时长

3 months

期刊介绍： The aim of Pharmacoepidemiology and Drug Safety is to provide an international forum for the communication and evaluation of data, methods and opinion in the discipline of pharmacoepidemiology. The Journal publishes peer-reviewed reports of original research, invited reviews and a variety of guest editorials and commentaries embracing scientific, medical, statistical, legal and economic aspects of pharmacoepidemiology and post-marketing surveillance of drug safety. Appropriate material in these categories may also be considered for publication as a Brief Report. Particular areas of interest include: design, analysis, results, and interpretation of studies looking at the benefit or safety of specific pharmaceuticals, biologics, or medical devices, including studies in pharmacovigilance, postmarketing surveillance, pharmacoeconomics, patient safety, molecular pharmacoepidemiology, or any other study within the broad field of pharmacoepidemiology; comparative effectiveness research relating to pharmaceuticals, biologics, and medical devices. Comparative effectiveness research is the generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat, and monitor a clinical condition, as these methods are truly used in the real world; methodologic contributions of relevance to pharmacoepidemiology, whether original contributions, reviews of existing methods, or tutorials for how to apply the methods of pharmacoepidemiology; assessments of harm versus benefit in drug therapy; patterns of drug utilization; relationships between pharmacoepidemiology and the formulation and interpretation of regulatory guidelines; evaluations of risk management plans and programmes relating to pharmaceuticals, biologics and medical devices.