Table 1. Risk of bias of included observational studies using ROBINS-I

Reference	Bias owing to confounding	Bias in selection of participants into the study	Bias inclassification of interventions	Bias owing todeviations from intended interventions	Bias owing to missing data	Bias inmeasurement of outcomes	Bias in selection of the reported results	Overall risk of bias
Uscher-Pines (US, 2016)²⁷	Serious	Moderate	Moderate	Serious	No available information	Moderate	No available information	Moderate
Gordon (US, 2017)¹⁶	Serious	Moderate	Moderate	Moderate	No available information	Moderate	No available information	Moderate
Shi (US, 2018)²⁵	Moderate	Serious	Moderate	Moderate	No available information	Moderate	No available information	Moderate
Davis (US, 2019)¹⁵	Serious	Moderate	Moderate	Serious	No available information	Moderate	No available information	Serious
Halpren-Ruder (US, 2019)¹⁷	Serious	Serious	Moderate	Moderate	No available information	Moderate	No available information	Moderate
Ray (US, 2019)²⁴	Moderate	Serious	Moderate	Moderate	No available information	Moderate	No available information	Moderate
Miller (US, 2020)¹⁹	Serious	Moderate	Moderate	Serious	No available information	Moderate	No available information	Moderate
Murray (US, 2020)²⁰	Serious	Serious	Moderate	Serious	No available information	Moderate	No available information	Serious
Penza (US, 2020 A)²²	Serious	Serious	Serious	Serious	No available information	Moderate	No available information	Serious
Penza (US, 2020 B)²³	Serious	Serious	Serious	Serious	No available information	Moderate	No available information	Serious
Stenehjem (US, 2020)²⁶	Serious	Serious	Moderate	Serious	Moderate	Moderate	No available information	Moderate
Norden (US, 2020)²¹	Serious	Serious	Serious	Serious	No available information	Moderate	No available information	Serious