Addressing Extreme Propensity Scores via the Overlap Weights
The DOI matches its Crossref record; 19 of the 19 references it cites were found in a registry or an index.
Abstract
The popular inverse probability weighting method in causal inference is often hampered by extreme propensity scores, resulting in biased estimates and excessive variance. A common remedy is to trim patients with extreme scores (i.e., remove them from the weighted analysis). However, such methods are often sensitive to the choice of cutoff points and discard a large proportion of the sample. The implications for bias and the precision of the treatment effect estimate are unclear. These problems are mitigated by a newly developed method, the overlap weighting method. Overlap weights emphasize the target population with the most overlap in observed characteristics between treatments, by continuously down-weighting the units in the tails of the propensity score distribution. Here we use simulations to compare overlap weights to standard inverse probability weighting with trimming, in terms of bias, variance, and 95% confidence interval coverage. A range of propensity score distributions are considered, including settings with substantial nonoverlap and extreme values. To facilitate practical implementation, we further provide a consistent estimator for the standard error of the treatment effect estimated using overlap weighting.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Overlap weighting emphasizes patients for whom either strategy was clinically plausible and limits the influence of extreme weights (17, 18).”
From Day-4 antipseudomonal and anti-MRSA de-escalation was not associated with higher mortality in critically ill patients with sepsis (2026) · cites this paper as [17]
Citations
Cites
The central role of the propensity score in observational studies for causal effects
ROSENBAUM, P. R., & RUBIN, D. B. (1983) · Biometrika · doi:10.1093/biomet/70.1.41 · Cited by 31,241
Matching Methods for Causal Inference: A Review and a Look Forward
Stuart, E. A. (2010) · Statistical Science · doi:10.1214/09-sts313 · Cited by 5,478
Austin, P. C., & Stuart, E. A. (2015) · Statistics in Medicine · doi:10.1002/sim.6607 · Cited by 4,495
Constructing Inverse Probability Weights for Marginal Structural Models
Cole, S. R., & Hernan, M. A. (2008) · American Journal of Epidemiology · doi:10.1093/aje/kwn164 · Cited by 2,796
Lunceford, J. K., & Davidian, M. (2004) · Statistics in Medicine · doi:10.1002/sim.1903 · Cited by 1,674
Hirano, K., & Imbens, G. W. (2001) · Health Services and Outcomes Research Methodology · doi:10.1023/a:1020371312283 · Cited by 1,039
Dealing with limited overlap in estimation of average treatment effects
Crump, R. K., Hotz, V. J., Imbens, G. W., & Mitnik, O. A. (2009) · Biometrika · doi:10.1093/biomet/asn055 · Cited by 1,007
Balancing Covariates via Propensity Score Weighting
Li, F., Morgan, K. L., & Zaslavsky, A. M. (2018) · Journal of the American Statistical Association · doi:10.1080/01621459.2016.1260466 · Cited by 962
Self-citationRothman, K. J. (1988) · Cited by 690
Weight Trimming and Propensity Score Weighting
Lee, B. K., Lessler, J., & Stuart, E. A. (2011) · PLoS ONE · doi:10.1371/journal.pone.0018174 · Cited by 500
Cited by
Items 401–500 of 1,000
Mokart, D., Serre, E., Bruneel, F., Kouatchet, A., Lemiale, V., Chow‐Chine, L., Faucher, M., Sannini, A., Valade, S., Bisbal, M., et al. (2024) · Thrombosis Research · doi:10.1016/j.thromres.2024.03.027 · Cited by 6
Kawauchi, A., Aoki, M., Kitamura, N., Tagami, T., Hayashida, K., Aso, S., Yasunaga, H., & Nakamura, M. (2024) · The American Journal of Emergency Medicine · doi:10.1016/j.ajem.2024.04.034 · Cited by 6
Ho, C.-T., Fu, C., Tan, E. C., Kao, W., Lee, P., Huang, Y., Huo, T., Hou, M., Wu, J., & Su, C. (2024) · Journal of Gastroenterology and Hepatology · doi:10.1111/jgh.16640 · Cited by 6
“41,43,44 In Taiwan, patients with chronic diseases were permitted to receive a re fi ll card for their prescriptions every three months (90 days), and each re fi ll prescription was capped at a maximum of 28 days under Taiwan ’ s NHI regulations.”
Xu, S., Zheng, B., Su, B., Finkelstein, S. N., Welsch, R. E., Ng, K., & Shahn, Z. (2024) · American Journal of Epidemiology · doi:10.1093/aje/kwae217 · Cited by 6
Nulsen, B., Jensen, D. M., Kovacs, T. O. G., Ghassemi, K. A., Kaneshiro, M., Dulai, G. S., Jutabha, R., & GORNBEIN, J. A. (2020) · Digestive Diseases and Sciences · doi:10.1007/s10620-020-06679-4 · Cited by 6
“The propensity score was computed via logistic regression on visual versus DEP treatment [17].”
MethodCosta, O. S., O’Donnell, B., Vardar, B., Abdelgawwad, K., Brescia, C. W., Sood, N. A., & Coleman, C. I. (2021) · Current Medical Research and Opinion · doi:10.1080/03007995.2021.1947217 · Cited by 6
“Propensity scoreoverlap weighted Cox proportional hazard regression with robust sandwich estimators were used to calculate hazard ratios (HR) with corresponding 95% confidence intervals (CI)[22].”
MethodYamazaki, R., Ohbe, H., Matsuda, Y., Kito, S., Shigeta, M., Morita, K., Matsui, H., Fushimi, K., & Yasunaga, H. (2022) · Journal of Affective Disorders · doi:10.1016/j.jad.2022.06.060 · Cited by 6
French, M. A., Hayes, H., Johnson, J. K., Young, D. L., Roemmich, R. T., & Raghavan, P. (2024) · medRxiv · doi:10.1101/2024.01.08.24301026 · Cited by 6
Decision-Centered Abstractions via Orthogonal Estimation of Difference-of-Q FunctionsPreprint
Cao, D., & Zhou, A. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2406.08697 · Cited by 6
“Note that the (single-stage) R-learner loss function is an overlap-weighted (Li et al., 2019) regression against the doubly-robust score (DR-learner (Kennedy, 2020)).”
BackgroundFederated Causal Inference from Multi-Site Observational Data via Propensity Score AggregationPreprint
Khellaf, R., Bellet, A., & Josse, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.17961 · Cited by 6
Ross, T. L., Na, R., Au‐Yeung, G., DeFazio, A., Friedlander, M. L., Sivakumaran, T., Livingstone, K., Nagle, C. M., O'Neill, H., Williams, M., et al. (2024) · Gynecologic Oncology · doi:10.1016/j.ygyno.2024.07.684 · Cited by 5
Rim, J. G., Hellkamp, A. S., Neely, M. L., Reynolds, J. M., Belperio, J. A., Budev, M. M., Eason, L., Frankel, C. W., Keshavjee, S. H., Kirchner, J., et al. (2024) · The Journal of Heart and Lung Transplantation · doi:10.1016/j.healun.2024.11.033 · Cited by 5
Zhong, Z., Wang, J., Xia, L., Jing, S., Lai, Y., Cai, T., & Yang, P. (2025) · Arthritis & Rheumatology · doi:10.1002/art.43382 · Cited by 5
“30 This means that by downweighting individuals with extreme propensity scores (those who are very likely or very unlikely to receive the treatment), covariate balance can be achieved and valid comparisons across treatment groups can be enabled even if the original sample sizes are unevenly…”
Variance Estimation for Weighted Average Treatment Effects
Li, H., Liu, Y., Zhou, Y., Liu, J., Fu, D., & Matsouaka, R. (2025) · Statistics in Biosciences · doi:10.1007/s12561-025-09503-7 · Cited by 5
“…weight function (Li et al., 2018)), which weights each observation according to how close it to the “clinical equipoise” (Thomas et al., 2020; Li et al., 2019; Matsouaka and Zhou, 2024), More generally, in this paper, we will consider the average causal effect defined on a weighted…”
A Tutorial for Propensity Score Weighting Methods Under Violations of the Positivity Assumption
Liu, Y., Wang, Y., Gao, Y., Poteat, T., & Matsouaka, R. (2025) · Statistics in Medicine · doi:10.1002/sim.70329 · Cited by 5
“The resulting trimmed subpopulation is defined as { X ∶ 𝛼 𝑒 ( X ) 1 − 𝛼 } and is equivalent to applying a tilting function ℎ ( X ) = I ( 𝛼 𝑒 ( X ) 1 − 𝛼 ) , where I ( ⋅ ) is the indicator function [5, 16, 17].”
Alkadri, J., Chen, M., Karkouti, K., Morais, S., Saskin, R., Grudzinski, A. L., Ouzounian, M., Callum, J. L., Lin, Y., McCluskey, S. A., et al. (2025) · BJA Open · doi:10.1016/j.bjao.2025.100508 · Cited by 5
“19 This results in smaller weights for extreme propensity score values, while patients with more common characteristics contribute more to estimation.”
Napolitano, A., Joensuu, H., Rothschild, S., Evans, D., Heinrich, M. C., Sutton, T. L., Mayo, S. C., Boye, K., Falkenhorst, J., Cesne, A. L., et al. (2026) · JAMA Oncology · doi:10.1001/jamaoncol.2026.0007 · Cited by 5
Statins and survival free of incident frailty among older US veterans
Qazi, S., CHAREST, B. R., Pajewski, N. M., Callahan, K. E., Houston, D. K., Espinoza, S. E., Forman, D. E., Alexander, K. P., Djoussé, L., Gaziano, J. M., et al. (2026) · European Heart Journal · doi:10.1093/eurheartj/ehag451 · Cited by 5
Argon, B. D., ÇAKIR, A., Çakır, S. D., Safalı, F., Elçioğlu, M. N., & Adaş, M. (2026) · Graefe s Archive for Clinical and Experimental Ophthalmology · doi:10.1007/s00417-026-07312-2 · Cited by 5
Lin, P.-J., Lee, T.-L., Hershey, M. S., Shih, C.-W., Tausi, S., Sosene, V., Maani, P. P., Tupulaga, M., Wu, S. M., López‐Gil, J. F., et al. (2023) · Journal of Global Health · doi:10.7189/jogh.13.04097 · Cited by 5
Effect of multidisciplinary care on diabetic kidney disease: a retrospective cohort study
Hayashi, A., Mizuno, K., Shinkawa, K., Sakoda, K., Yoshida, S., Takeuchi, M., Yanagita, M., & Kawakami, K. (2024) · BMC Nephrology · doi:10.1186/s12882-024-03550-w · Cited by 5
“Fifth, we conducted inverse probability of treatment weighting (IPTW) and overlap weighting [35] instead of PS matching.”
MethodWalter, J., Ma, J. E., Platt, A. C., Acker, Y., Sendak, M., Gao, M. C., Gardner, M., Balu, S., & Setji, N. (2024) · Journal of Brown Hospital Medicine · doi:10.56305/001c.120907 · Cited by 5
Asada, T., Simon, C. Z., Durbas, A., Allen, M. R. J., DiSilvestro, K. J., Hirase, T., Singh, N., Bovonratwet, P., Heuer, A., Subramanian, T., et al. (2024) · The Spine Journal · doi:10.1016/j.spinee.2024.10.007 · Cited by 5
The Effect of Peer Instruction Lectures on Learning Attitudes in Epidemiology Education.
Mitsuhashi, T. (2021) · Institutional Repositories DataBase (IRDB) · doi:10.18926/amo/62773 · Cited by 5
Syrogiannouli, L., Wildisen, L., Meuwese, C. L., Bauer, D. C., Cappola, A. R., Gussekloo, J., den Elzen, W. P. J., Trompet, S., Westendorp, R. G. J., Jukema, J. W., et al. (2022) · Frontiers in Psychiatry · doi:10.3389/fpsyt.2022.774251 · Cited by 5
“One reason that may explain why the propensity score analysis does not generally agree with the ANCOVA analysis is the imbalance exposure “allocation ratio” that may produce a lack of overlap in the estimated propensity score by exposure groups and consequent extreme weights (47).”
BackgroundInstrumental variable estimation of weighted local average treatment effects
Choi, B. Y. (2023) · Statistical Papers · doi:10.1007/s00362-023-01415-2 · Cited by 5
Feeney, C. D., Chandler, M., Platt, A. C., Sun, S., Setji, N., & Ming, D. Y. (2023) · Journal of Hospital Medicine · doi:10.1002/jhm.13234 · Cited by 5
Jang, S., Choi, H. G., Kwon, M. J., Kim, J. H., Kim, J., Kim, Y., & Kim, S. Y. (2024) · Biomedicines · doi:10.3390/biomedicines12010170 · Cited by 5
“The propensity score (PS) was applied and overlap weighting was calculated [22 The variables of study and control groups were compared using the standardized di ff erence.”
MethodKewcharoen, J., Chang, A. K., Parwani, P. J., Fraser, G. E., Bharadwaj, A. S., Seliem, A., Tran, D. D., Stoletniy, L. N., Sakr, A. E., & Abramov, D. (2022) · Cardiology Research · doi:10.14740/cr1364 · Cited by 5
“This was demonstrated with Love plots in our data and analyses [10].”
MethodMyers, L. C., Kipnis, P., Greene, J., Chen, A., Creekmur, B., Xu, S., Sankar, V., Roubinian, N. H., Langer-Gould, A. M., Gould, M. K., et al. (2023) · Journal of Critical Care · doi:10.1016/j.jcrc.2023.154322 · Cited by 5
“[26-28] We also performed two sensitivity analyses using inverse probability of treatment weighting.”
Propensity score analysis for health care disparities: a deweighting approach
Choi, B. Y. (2024) · BMC Medical Research Methodology · doi:10.1186/s12874-024-02230-5 · Cited by 5
“(cid:31) (cid:30) The overlap weights are useful to address extreme PSs [11].”
BackgroundZhang, J., Sum, S.-Y., Hsu, J.-G., Chiang, M.-F., Lee, T., & Wu, S. (2021) · Journal of Personalized Medicine · doi:10.3390/jpm11121358 · Cited by 5
Iwai, C., Miyawaki, A., Konishi, T., Okada, A., Fujita, A., Jo, T., & Yasunaga, H. (2025) · Breast Cancer Research and Treatment · doi:10.1007/s10549-025-07720-8 · Cited by 4
A Bayesian Approach to the G‐Formula via Iterative Conditional Regression
Liu, R., Hu, L., Wilson, F. P., Warren, J. L., & Li, F. (2025) · Statistics in Medicine · doi:10.1002/sim.70123 · Cited by 4
“While various strategies have been proposed to address positivity violations in the context of time-fixed treatments [69, 70], their extensions to time-varying treatments remain less developed.”
Self-citationCale, J., Whitten, T., Perales, F., O’Shannessy, D. M., & Leclerc, B. (2025) · Journal of Criminal Justice · doi:10.1016/j.jcrimjus.2025.102462 · Cited by 4
Choi, B. Y., & Brookhart, M. A. (2025) · Pharmacoepidemiology and Drug Safety · doi:10.1002/pds.70173 · Cited by 4
Lin, C., Chen, Y., Chang, K., & Chi, P. (2025) · Journal of Asian Economics · doi:10.1016/j.asieco.2025.101996 · Cited by 4
Sansom, K., Mittinty, M., Dunbar, C., Vakulin, A., Adams, R. J., McArdle, N., Eastwood, P. R., & Reynolds, A. C. (2025) · Sleep Medicine · doi:10.1016/j.sleep.2025.106699 · Cited by 4
Wee, L. E., Ho, R. W. L., Lim, J. T., Chiew, C. J., Ong, B., Said, Z., Lim, S. K., Young, B. E., Lye, D. C., & Tan, K. B. (2025) · Clinical Microbiology and Infection · doi:10.1016/j.cmi.2025.10.013 · Cited by 4
Wang, K., Rosman, L., & Lu, H. (2026) · European Journal of Epidemiology · doi:10.1007/s10654-025-01341-7 · Cited by 5
Lo, M. Y., Amin, A. K., Kowalski, M., Kroman, A. M., Mathew, D. J., Parikh, V. K., Weiner, S., Joseph, J., Ganz, L. I., & Mortada, M. E. (2026) · Journal of Cardiovascular Electrophysiology · doi:10.1111/jce.70255 · Cited by 4
“Overlap weights were calculated from propensity scores based on methods described by Li et al., which emphasize patients with the most overlap in their observed characteristics, providing more precise treatment effects [8].”
McCormick, N., Burrack, N., Yokose, C., Lu, N., Wexler, D. J., Rai, S. K., Challener, G. J., Aviña-Zubieta, J. A., McCoy, R. G., Abuhasira, R., et al. (2026) · Diabetes Care · doi:10.2337/dc25-1713 · Cited by 4
Grory, B. M., Preterre, C., Lebranchu, P., Ryan, S. J., Jørstad, Ø. K., Moe, M. C., Tünnerhoff, J., Spitzer, M. S., Grohmann, C., Dumitrascu, O. M., et al. (2026) · PLoS ONE · doi:10.1371/journal.pone.0342739 · Cited by 4
Kwo, P. Y., Telep, L. E., Hong, L., Chuo, C., Chiang, B., Naik, S., Frenette, C. T., Singer, A. W., Chokkalingam, A. P., Wong, R. J., et al. (2026) · Alimentary Pharmacology & Therapeutics · doi:10.1111/apt.70613 · Cited by 4
“For the mid-level and high-level viraemia sensitivity analyses, overlap weights were calculated to balance the cohorts [17].”
Lee, S. B., Park, C. M., Levin, R., & Kim, D. H. (2024) · Journal of the American Geriatrics Society · doi:10.1111/jgs.19067 · Cited by 4
“21 The overlap weighting estimates the average treatment effect in a population with the most overlap in clinical characteristics across the treatment groups and avoids extreme weights.”
Lim, Y., Jang, J.-H., Lee, S. H., Ahn, J. H., Hong, Y. J., Ahn, Y., Jeong, M. H., Kim, C. J., Hahn, J., Lee, J. M., et al. (2024) · Frontiers in Cardiovascular Medicine · doi:10.3389/fcvm.2024.1475483 · Cited by 4
Assessing racial disparities in healthcare expenditure using generalized propensity score weighting
Liu, J., Liu, Y., Zhou, Y., & Matsouaka, R. (2025) · BMC Medical Research Methodology · doi:10.1186/s12874-025-02508-2 · Cited by 4
“We need to point out that ATT estimation through both trimming and truncation is subjective since it involves ad hoc selections of a user-specified threshold parameter(s) α . this often leads to loss of information and sensitivity of the estimated racial disparities [23, 25, 39].”
InfluentialBraïk, R., Berna, P., Walbecq, J., Germain, Y., Flet, T., & Arab, O. A. (2025) · Scientific Reports · doi:10.1038/s41598-025-16171-6 · Cited by 4
“We applied overlap weighting to estimate the average treatment effect within the overlap population, as described by Fan Li et al. 11 .”
Generalizing Causal Effects with Noncompliance: Application to Deep Canvassing Experiments
Chen, Z., & Huang, M. (2026) · Political Analysis · doi:10.1017/pan.2026.10063 · Cited by 4
Evaluating sociodemographic bias in a deployed machine-learned patient deterioration model
Colacci, M., Pou-Prom, C., Siddiqi, A., Mamdani, M., & Verma, A. A. (2025) · JAMIA Open · doi:10.1093/jamiaopen/ooaf158 · Cited by 4
“29,30 Weighting was completed separately for the overall cohort, as well as within each sociodemographic subgroup.”
Thyroid Cancer and Psoriasis: A Nested Case–Control Study
Kim, S. Y., Yoo, D. M., Chung, J., & Choi, H. G. (2022) · Diagnostics · doi:10.3390/diagnostics12102297 · Cited by 4
Counterfactual Balancing Feature Alignment for Few-Shot Cross-Domain Scene Parsing
Sun, C. M., Chen, Z., & Lin, L. (2023) · IEEE Access · doi:10.1109/access.2023.3263561 · Cited by 4
“There have been many approaches invented to balance the covariates across distinct groups for causal estimation, where propensity score [20], [36] and counterfactual matching [25] are two categories that have been mostly adopted.”
Lusk, J. B., Wilson, L. E., Nalwade, V., Song, A., Schrag, M., Biousse, V., Fan, L., Poli, S., Piccini, J. P., Xian, Y., et al. (2023) · PLoS ONE · doi:10.1371/journal.pone.0296251 · Cited by 4
Laici, C., Gamberini, L., Allegri, D., Bianchini, A., Vitale, G., Losito, M., Morini, L., Prosperi, E., Ravaioli, M., Cescon, M., et al. (2024) · Internal and Emergency Medicine · doi:10.1007/s11739-024-03530-w · Cited by 4
“The ATO is the effect of the treatment in an equipoise population, in practice, a population of eligible patients for whom either treatment is currently equally implemented or there is no strong preference for one over the other [19].”
Ren, A., Fan, M., Gu, Z., Liang, X., Xu, L., Liu, C., Wang, D., Chang, H., & Zhu, M. (2024) · British Journal of Clinical Pharmacology · doi:10.1111/bcp.16056 · Cited by 4
“19 The balance before and after applying the weights was investigated using the standardized mean difference (SMD), where an SMD 0.1 was considered a good balance.”
MethodInfluentialLian, X., Bai, Y., Du, P., Jing, Z., Gao, J., Liu, F., Hu, J., & Xi, Y. (2024) · The Journal of Steroid Biochemistry and Molecular Biology · doi:10.1016/j.jsbmb.2024.106629 · Cited by 4
Wee, L. E., Lim, J. T., Tan, J. Y. J., Chiew, C. J., Yung, C.-F., Chong, C. Y., Lye, D. C., & Tan, K. B. (2025) · Open Forum Infectious Diseases · doi:10.1093/ofid/ofaf134 · Cited by 4
“…in baseline characteristics between the COVID-19 and dengue cases were adjusted by overlap weighting to better account for outlying values [32, 33], with overlap weights computed as equal to the propensity score for individuals infected with SARS-CoV-2 (the controls) and 1 – propensity…”
Iwai, C., Jo, T., Okada, A., Fujita, A., Konishi, T., Oba, K., Hashimoto, Y., & Yasunaga, H. (2025) · Canadian Journal of Ophthalmology · doi:10.1016/j.jcjo.2025.02.010 · Cited by 4
Yajima, W., Aso, S., Matsui, H., Fushimi, K., & Yasunaga, H. (2025) · Journal of Intensive Care · doi:10.1186/s40560-025-00788-w · Cited by 4
“To adjust for confounders, we conducted a propen-sity-score overlap weighting [15–17].”
Remdesivir and risk of long-term multi-systemic sequelae following COVID-19 hospitalization
Wee, L. E., Lim, J. T., Tay, A. T., Chiew, C. J., Young, B. E., Vasoo, S. S., Lou, H. X., Lye, D. C., & Tan, K. B. (2025) · Clinical Microbiology and Infection · doi:10.1016/j.cmi.2025.06.016 · Cited by 4
Madaras‐Kelly, K., Boyd, J. D., & Bond, L. (2025) · Medicine · doi:10.1097/md.0000000000043172 · Cited by 4
“[36,37] These weights were applied to each patient visit with the goal of balancing the 2 patient groups (e.g., telehealth, primary care) with respect to the potential confounders (e.g., why patients were seen in telehealth or primary care).”
Yang, I., Chong, K.-S., Peng, Z., Ou, H., & Wang, M.-C. (2025) · Diabetes Research and Clinical Practice · doi:10.1016/j.diabres.2025.112427 · Cited by 4
Tong, G., Li, F., & Allen, A. S. (2022) · Springer eBooks · doi:10.1007/978-3-319-52636-2_117 · Cited by 4
Self-citationRubens, F. D., Fremes, S. E., Grubic, N., Fergusson, D. A., Taljaard, M., & van Walraven, C. (2023) · Journal of Thoracic and Cardiovascular Surgery · doi:10.1016/j.jtcvs.2023.02.008 · Cited by 4
Kwon, M. J., Kim, J.-K., Kim, M.-J., Yoo, D. M., Lee, N.-E., Han, K. M., Kim, N. Y., Kang, H. S., Choi, H. G., & Kim, E. S. (2024) · Journal of Personalized Medicine · doi:10.3390/jpm14040356 · Cited by 4
Impact of statin use in head and neck cancer: a nested case–control study
Rhi, H., Choi, H. G., & Kang, J. W. (2024) · European Archives of Oto-Rhino-Laryngology · doi:10.1007/s00405-024-09020-1 · Cited by 4
Jemielita, T., Li, X. N., Burke, T. A., Liaw, K., Zhou, W., & Chen, C. (2021) · JCO Clinical Cancer Informatics · doi:10.1200/cci.21.00042 · Cited by 4
Ke, S., Wang, W., Li, B., Feng, X., Yan, D., & Liu, J. (2023) · Frontiers in Oncology · doi:10.3389/fonc.2022.1032063 · Cited by 4
“The OW method overcomes essential limitations of traditional weighting approaches by emphasizing the target population with the most overlap (42, 43).”
MethodKwon, M. J., Han, K. M., Kim, J., Kim, J. H., Kim, M.-J., Kim, N. Y., Choi, H. G., & Kang, H. S. (2023) · Cancers · doi:10.3390/cancers15235606 · Cited by 4
“Overlap weighting values range from 0 to 1 and are designed to achieve exact balance while optimizing the precision [53,54].”
Voza, A., Desai, A., Luzzi, S., Lucifero, A. G., Azzolini, E., Kogan, M., Goretti, G., Piovani, D., Bonovas, S., Angelotti, G., et al. (2021) · Applied Sciences · doi:10.3390/app11199342 · Cited by 4
“Considering that age and underlying comorbidities are strongly associated with mortality and LOS, and that they were unbalanced between the first and second waves of COVID-19 patients, an overlap propensity score weighting was performed to address potential confounding factors [18,19].”
BackgroundMethodThe concomitant use of proton pump inhibitors and BRAF/MEK inhibitors in metastatic melanoma
Poizeau, F., Balusson, F., Lemaitre, F., Tron, C., Pracht, M., Russo, D., Dinulescu, M., Lesimple, T., Oger, E., & Dupuy, A. (2023) · British Journal of Dermatology · doi:10.1093/bjd/ljac085 · Cited by 4
Propensity Score Estimation Using Density Ratio Model under Item NonresponsePreprint
Wang, H., & Kim, J. K. (2021) · arXiv (Cornell University) · Cited by 4
“Double Machine Learning for Causal Inference in High-Dimensional Electronic Health Records”Preprint
Du, M., Guo, Y., Li, X., Català, M., & Prieto‐Alhambra, D. (2025) · medRxiv · doi:10.1101/2025.07.21.25331944 · Cited by 4
Jiang, C., Thompson, M., & Wallace, M. P. (2024) · Statistical Methods in Medical Research · doi:10.1177/09622802241242313 · Cited by 3
Takeda, K., Yokoyama, A., Fukami, T., Kimura, Y., Suzukawa, M., Jo, T., Suzuki, J., Sasaki, Y., Mitani, A., Tanaka, G., et al. (2024) · Medical Mycology · doi:10.1093/mmy/myae117 · Cited by 3
Zarama, V., Vesga, C. E., Balanta-Silva, J., Barbosa, M. M., Quintero, J., Clarete, A., Vesga-Reyes, P. A., & Godinez, J. C. S. (2025) · Echo Research and Practice · doi:10.1186/s44156-025-00071-6 · Cited by 3
“This method also enhances the efficiency and precision of the estimated treatment effects by focusing on the population where the treatment effect is most identifiable [25–27].”
Yao, T., Chang, S., Tsai, Y., Chiang, S.-J., & Tsai, H. (2025) · Clinical and Translational Science · doi:10.1111/cts.70311 · Cited by 3
“We employed overlap weighting, a propensity- score-based approach, to balance baseline characteristics and minimize potential confounding effects [20].”
Health Service Use Among Young Adults With a History of Adolescent Cannabis Use
Martínez, P., Chadi, N., Ryan, N. C., Vergunst, F., Dorais, M., Séguin, J. R., Vitaro, F., Temcheff, C. E., Tremblay, R. E., Boivin, M., et al. (2025) · JAMA Network Open · doi:10.1001/jamanetworkopen.2025.39977 · Cited by 3
Overlap Weights for Binary Outcomes: A Performance Assessment
Park, S. Y., Ahn, J., Lee, J. H., Kwon, J., & Lee, H. (2025) · Pharmacoepidemiology and Drug Safety · doi:10.1002/pds.70253 · Cited by 3
“These theoretical properties of OW have been confirmed in simulation studies with a continuous outcome setting, demonstrating that OW outperforms IPW in terms of bias, variance, and coverage probability in data with limited overlap [8, 10, 11].”
InfluentialZhang, B., O’Brien, S. M., Wu, Y. S., & Thomas, L. E. (2026) · Statistics in Medicine · doi:10.1002/sim.70643 · Cited by 3
Self-citationYu, Y., Jiao, Z., Wu, T., Ning, K., Cai, T., Yang, Z., Chen, W., Jiang, M. J., & Yang, A. (2026) · Cancer Immunology Immunotherapy · doi:10.1007/s00262-025-04284-w · Cited by 3
Suzuki, T., Yamamoto, N., Miura, T., Otaira, Y., Fujiwara, S., Yamashita, T., & Nakajima, T. (2026) · Bone & Joint Open · doi:10.1302/2633-1462.72.bjo-2025-0352.r1 · Cited by 3
“30 Propensity scores were estimated by logistic regression using the following covari ates: age, sex, BMI, diagnosis, laterality, surgical approach, and pelvic tilt angle.”
Madaras‐Kelly, K., Boyd, J. S., & Bond, L. (2026) · Journal of Comparative Effectiveness Research · doi:10.57264/cer-2025-0163 · Cited by 3
Brownell, N. K., Solomon, N. C., Greene, S. J., Chiswell, K., Vaduganathan, M., Yancy, C. W., Hsue, P. Y., Ziaeian, B., & Fonarow, G. C. (2026) · Journal of the American Heart Association · doi:10.1161/jaha.126.049161 · Cited by 3
Shu, S., Hamasaki, T., Evans, S., Komarow, L., van Duin, D., & Diao, G. (2026) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2026.2697866 · Cited by 3
Moreno, A., Vargas, C. O., Azocar, F., Villarroel‐Espíndola, F., Cofré, M., Oppliger, H., Ríos, F., Raijmakers, M., Silva-Ayarza, I., Beltrán, C., et al. (2021) · International Journal of Infectious Diseases · doi:10.1016/j.ijid.2021.09.038 · Cited by 3
“…difference in treatment effect between the two roups, the average treatment in the overlap population (ATO) was sed with the overlap model and the average treatment effect ATE) was used in the IPW model; both were estimated using the package PSweight ( Li and Thomas, 2019 ; Thomas et al., 2020 ).”
Association between the Use of Statins and Brain Tumors
Jang, S., Choi, H. G., Kwon, M. J., Kim, J. H., Kim, J., & Kim, S. Y. (2023) · Biomedicines · doi:10.3390/biomedicines11082247 · Cited by 3
Platelet Reduction after Aortic Bioprosthesis Implantation: Results from the PORTRAIT Study
Jiritano, F., Serraino, G. F., Mauro, M. D., Borelli, M., Scrofani, R., Patanè, L., Caporali, E., Matteucci, M., Fina, D., Kowalewski, M., et al. (2023) · Journal of Clinical Medicine · doi:10.3390/jcm12237414 · Cited by 3
Xu, H., Xing, Z., Ai, K., Wang, J., Lv, Z., Deng, H., Li, K., Wang, Y., & Li, Y. (2024) · World Journal of Surgical Oncology · doi:10.1186/s12957-024-03302-y · Cited by 3
“In the four subsets cohort we used overlap weighting (OW) [12] to balance the confounding factors caused by the nonrand-omized design of this study.”
MethodYouk, H., Lee, H. Y., Lee, E. Y., Kim, Y. J., Park, J. Y., Choi, H. G., Kim, H. S., & Lee, J. W. (2025) · BMC Musculoskeletal Disorders · doi:10.1186/s12891-025-08527-w · Cited by 3
“This method generated weights ranging from 0 to 1, achieving exact balance and enhancing statistical precision [16–18].”
Elbaz, A., Pesquet, C., Defaye, M., Delignette, M., Despres, C., Sakek, R., Weiss, E., Bignon, A. L., Devauchelle, P., & Dewitte, A. (2025) · Transplantation · doi:10.1097/tp.0000000000005442 · Cited by 3
Li, H., Lindberg, V., Zhu, L., Huang, X., Feng, J., & Baak, J. P. A. (2025) · Integrative Cancer Therapies · doi:10.1177/15347354251342739 · Cited by 3
“25-28 Matching (1:1 greedy algorithm) and IPTW were used to adjust for baseline covariate imbalance 29,30 (Supplemental Table 1).”
Harrison, O. J., Tan, K. S., Dycoco, J., Gray, K. D., Sihag, S., Molena, D., Bott, M., Rocco, G., Isbell, J. M., Adusumilli, P. S., et al. (2025) · JTCVS Techniques · doi:10.1016/j.xjtc.2025.06.007 · Cited by 3
Sumida, K., Mallisetty, Y., Shrestha, P., Vo, A., Khan, R. N., Thomas, F., Kalantar‐Zadeh, K., & Kovesdy, C. P. (2025) · Clinical Journal of the American Society of Nephrology · doi:10.2215/cjn.0000000779 · Cited by 3
Wyss, R., Yang, J., Schneeweiß, S., Plasek, J. M., Zhou, L., DeRamus, T. P., Weberpals, J. G., Ngan, K., Tsacogianis, T., & Lin, K. J. (2025) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2025.104882 · Cited by 3
Sasaki, A., Nakajima, M., Shinozaki, T., Sasabuchi, Y., Ohbe, H., Kaszynski, R. H., Kimura, Y., Morita, K., Goto, T., Aiyama, Y., et al. (2025) · Journal of Intensive Care · doi:10.1186/s40560-025-00826-7 · Cited by 3
“…weighting emphasizes the subpopulation with greater clinical equipoise (i.e., those with estimated propensity scores closer to 0.5), thereby shifting the target population from the overall cohort [32], we selected it as the primary analysis to take advantage of its improved statistical precision.”
InfluentialIwai, C., Konishi, T., Miyawaki, A., Okada, A., Isogai, T., Jo, T., & Yasunaga, H. (2025) · Breast Cancer Research and Treatment · doi:10.1007/s10549-025-07852-x · Cited by 3
“Overlap weighting mathematically ensures perfect balance, reducing all standardized differences to zero [23].”
Transcatheter Aortic Valve Replacement in Kidney Transplant Patients
Mentias, A. G., Desai, M. Y., Saad, M. S., Rossen, J. D., Megally, M., Jneid, H. M., Horwitz, P. A., & Vaughan‐Sarrazin, M. (2020) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2020.09.581 · Cited by 3
Caroti, K. S., Khorana, A. A., Becattini, C., Lee, A. Y. Y., Ekbom, A., Carrier, M., Cohen, A. T., Brescia, C. W., Abdelgawwad, K., Psaroudakis, G., et al. (2023) · Clinical and Applied Thrombosis/Hemostasis · doi:10.1177/10760296231189282 · Cited by 3
“0 (a concern with inverse probability weighting).(20) Our use of an intent-to-treat methodology may have resulted in bias toward a null effect, especially with longer follow-up durations.”
MethodInfluentialCamp, J., Bayrhuber, M., Anka, N., Heine, V., Glattacker, M., Farin‐Glattacker, E., & Rieg, S. (2023) · Infection · doi:10.1007/s15010-023-02088-7 · Cited by 3
“Secondly, overlap weights in contrast to other frequently used methods, e.g., propensity score matching, show less bias and seem more efficient [20].”
Method
Metadata
- Authors
- Li, F., & Thomas, L. E.
- Affiliations
- Duke University (US)
- Corresponding author
- Li, Fan
- Venue
- American Journal of Epidemiology
- Publisher
- Oxford University Press (OUP)
- Published
- 2018-09-05
- Volume
- 188
- Issue
- 1
- Pages
- 189-190
- ISSN
- 0002-9262 (ISSN-L)
- 1476-6256
- OpenAlex
- W2907932858
- Cited by
- 694
- References
- 19
- Citations by year
- 2026: 97 · 2025: 124 · 2024: 145 · 2023: 119 · 2022: 83 · 2021: 73 · 2020: 35 · 2019: 15 · 2018: 1
- FWCI
- 25.73
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model (am)
- Locations
- American Journal of Epidemiology (publishedVersion)
- PubMed (publishedVersion)
- Funders
- Agency for Healthcare Research and Quality (SES-1424688)
- National Science Foundation
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Clinical Trials
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Keywords
- inverse probability weighting, treatment effect estimation, bias and variance, confidence interval coverage, standard error estimation, causal inference
- MeSH terms
- Epidemiologic Methods, Humans, Models, Statistical, Bias, Causality, Propensity Score
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.