Quasi-rerandomization for observational studies

BMC Medical Research Methodology

Research Article

Wen Su¹  Guosheng Yin¹  Hengtao Zhang¹ 

[1] Department of Statistics and Actuarial Science, The University of Hong Kong, Hong Kong, China;

关键词: Causal inference;    Covariate balance;    Observational data;    Rerandomization;    Treatment effect;   

DOI  :  10.1186/s12874-023-01977-7

 received in 2022-10-28, accepted in 2023-06-13,  发布年份 2023

来源: Springer

【 摘 要 】

BackgroundIn the causal analysis of observational studies, covariates should be carefully balanced to approximate a randomized experiment. Numerous covariate balancing methods have been proposed for this purpose. However, it is often unclear what type of randomized experiments the balancing approaches aim to approximate; and this may cause ambiguity and hamper the synthesis of balancing characteristics within randomized experiments.MethodsRandomized experiments based on rerandomization, known for significant improvement on covariate balance, have recently gained attention in the literature, but no attempt has been made to integrate this scheme into observational studies for improving covariate balance. Motivated by the above concerns, we propose quasi-rerandomization, a novel reweighting method, where observational covariates are rerandomized to be the anchor for reweighting such that the balanced covariates obtained from rerandomization can be reconstructed by the weighted data.ResultsThrough extensive numerical studies, not only does our approach demonstrate similar covariate balance and comparable estimation precision of treatment effect to rerandomization in many situations, but it also exhibits advantages over other balancing techniques in inferring the treatment effect.ConclusionOur quasi-rerandomization method can approximate the rerandomized experiments well in terms of improving the covariate balance and the precision of treatment effect estimation. Furthermore, our approach shows competitive performance compared with other weighting and matching methods. The codes for the numerical studies are available at https://github.com/BobZhangHT/QReR.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

Fig. 1

41239_2023_389_Article_IEq76.gif

13015_2023_227_Article_IEq341.gif

Fig. 1

Fig. 1

Fig. 5

41239_2023_389_Article_IEq72.gif

Fig. 4

Fig. 5

41239_2023_389_Article_IEq68.gif

Fig. 3

Fig. 1

Fig. 8

Fig. 3

40691_2023_344_Figd_HTML.gif

Fig. 3

Fig. 4

Fig. 2

Fig. 9

Fig. 4

41239_2023_389_Article_IEq56.gif

Fig. 8

Fig. 2

Fig. 2

Fig. 1

Fig. 3

Fig. 5

Fig. 4

Fig. 2

Fig. 2

Fig. 4

Fig. 1

Fig. 3

Fig. 5

Fig. 1

Fig. 2

Fig. 4

40798_2023_601_Article_IEq8.gif

Fig. 5

Fig. 3

40798_2023_601_Article_IEq5.gif

40798_2023_601_Article_IEq3.gif

40798_2023_601_Article_IEq2.gif

Fig. 4

Fig. 3

Fig. 2

Fig. 3

Fig. 1

Fig. 3

Fig. 1

Fig. 3

Fig. 4

41239_2023_389_Article_IEq13.gif

Fig. 1

Fig. 1

Fig. 3

Fig. 2

41239_2023_389_Article_IEq8.gif

Fig. 2

Fig. 1

40708_2023_197_Article_IEq86.gif

40708_2023_197_Article_IEq84.gif

Fig. 1

40708_2023_197_Article_IEq81.gif

40708_2023_197_Article_IEq83.gif

40708_2023_197_Article_IEq82.gif

Fig. 4

40708_2023_197_Article_IEq79.gif

40708_2023_197_Article_IEq78.gif

Fig. 7

Fig. 2

40708_2023_197_Article_IEq71.gif

40708_2023_197_Article_IEq70.gif

Fig. 1

40708_2023_197_Article_IEq68.gif

Fig. 2

Fig. 4

40708_2023_197_Article_IEq62.gif

Fig. 1

40708_2023_197_Article_IEq60.gif

40708_2023_197_Article_IEq59.gif

Fig. 1

40708_2023_197_Article_IEq55.gif

40708_2023_197_Article_IEq54.gif

Fig.9

Fig. 5

40708_2023_197_Article_IEq51.gif

40708_2023_197_Article_IEq50.gif

Fig. 2

40708_2023_197_Article_IEq48.gif

40708_2023_197_Article_IEq47.gif

40708_2023_197_Article_IEq46.gif

Fig. 7

Fig. 1

Fig. 2

Fig. 1

Fig.8

Fig. 1

Fig. 1

40708_2023_197_Article_IEq35.gif

40798_2023_598_Article_IEq30.gif

40798_2023_598_Article_IEq29.gif

40798_2023_598_Article_IEq28.gif

Fig. 1

Fig. 2

40798_2023_598_Article_IEq26.gif

40798_2023_598_Article_IEq25.gif

40798_2023_598_Article_IEq24.gif

40798_2023_598_Article_IEq23.gif

40708_2023_197_Article_IEq25.gif

40798_2023_598_Article_IEq22.gif

40798_2023_598_Article_IEq21.gif

Fig. 1

40798_2023_598_Article_IEq19.gif

40798_2023_598_Article_IEq18.gif

40798_2023_598_Article_IEq17.gif

40798_2023_598_Article_IEq16.gif

40798_2023_598_Article_IEq15.gif

40798_2023_598_Article_IEq14.gif

Fig. 2

40798_2023_598_Article_IEq12.gif

Fig. 1

40798_2023_598_Article_IEq10.gif

40798_2023_598_Article_IEq9.gif

40798_2023_598_Article_IEq8.gif

40798_2023_598_Article_IEq7.gif

40798_2023_598_Article_IEq6.gif

40798_2023_598_Article_IEq5.gif

Fig. 14

40798_2023_598_Article_IEq2.gif

40798_2023_598_Article_IEq1.gif

Fig. 5

Fig. 4

40708_2023_197_Article_IEq1.gif

Fig. 3

Fig.6

Fig. 2

Fig. 5

Fig. 13

Fig. 4

Fig. 2

Fig.5

Fig. 7

12938_2023_1124_Article_IEq44.gif

Fig. 3

Fig. 1

Fig. 6

Fig. 2

Fig. 12

Fig. 4

Fig. 8

Fig. 7

42004_2023_934_Article_IEq6.gif

Fig. 3

Fig. 7

Fig. 2

Fig. 2

42004_2023_934_Article_IEq2.gif

Fig. 7

12938_2023_1124_Article_IEq11.gif

12938_2023_1124_Article_IEq13.gif

Fig. 1

Fig. 6

12938_2023_1124_Article_IEq9.gif

Fig. 1

Fig. 5

Fig. 3

Fig. 1

Fig.3

Fig. 2

41534_2023_739_Article_IEq82.gif

Fig. 7

Fig. 2

Fig. 1

Fig. 5

Fig.2

Fig. 4

Fig. 1

Fig. 2

Fig. 4

Fig. 11

Fig. 6

Fig. 3

Fig. 1

Fig. 4

42004_2023_944_Article_IEq9.gif

Fig.1

Fig. 6

42004_2023_944_Article_IEq7.gif

Fig. 4

Fig. 1

Fig. 3

40854_2023_486_Article_IEq61.gif

Fig. 2

40854_2023_486_Article_IEq60.gif

Fig. 5

Fig. 7

40854_2023_486_Article_IEq57.gif

Fig. 5

40854_2023_486_Article_IEq55.gif

42004_2023_944_Article_IEq6.gif

Fig. 5

42004_2023_944_Article_IEq4.gif

40854_2023_486_Article_IEq48.gif

Fig. 3

40854_2023_486_Article_IEq45.gif

Fig. 3 (abstract A15)

40854_2023_486_Article_IEq44.gif

Fig. 2 (abstract A15)

40854_2023_486_Article_IEq42.gif

Fig. 1 (abstract A15)

40854_2023_486_Article_IEq39.gif

40854_2023_486_Article_IEq38.gif

Fig. 6

40854_2023_486_Article_IEq35.gif

Fig. 1 (abstract A14)

40854_2023_486_Article_IEq33.gif

Fig. 1

40854_2023_486_Article_IEq30.gif

Fig. 1

Fig. 1

40854_2023_486_Article_IEq28.gif

40854_2023_486_Article_IEq27.gif

40854_2023_486_Article_IEq26.gif

Fig. 4

40854_2023_486_Article_IEq72.gif

Fig. 3

13690_2023_1118_Article_IEq7.gif

13690_2023_1118_Article_IEq6.gif

13690_2023_1118_Article_IEq5.gif

13690_2023_1118_Article_IEq4.gif

13690_2023_1118_Article_IEq3.gif

13690_2023_1118_Article_IEq2.gif

13690_2023_1118_Article_IEq1.gif

Fig. 8

Fig. 3

40854_2023_486_Article_IEq17.gif

Fig. 8

Fig. 2

Fig. 1

Fig. 5

Fig. 3

Fig. 7

Fig. 3

Fig. 4

Fig. 7

Fig. 2

Fig. 6

Fig. 5

Fig. 3

Fig. 4

Fig. 2

Fig. 7

12876_2023_2855_Article_IEq1.gif

Fig. 2

Fig. 6

Fig. 9

Fig. 1

Fig. 5

Fig. 6

Fig. 2

Fig. 4

Fig. 3

Fig. 1

Fig. 2

Fig. 3

Fig. 2

Fig. 2

Fig. 2

Fig. 1

Fig. 2

Fig. 1

Fig. 1

Fig. 1

Fig. 1

Fig. 1

Fig. 2

Fig. 5

Fig. 2

Fig. 1

Fig. 3