Table 1.

Minimum sufficient adjusted sets (MSAS) for each predictor examined in each regression models.

Predictor variables	MSAS*
Age	N/A
Gender	N/A
Pre-COVID mental health	Age, occupation, gender
Physical health	Age, COVID diagnosis, Pre-COVID mental health, occupation, gender, supervisor support
Occupation	Age, gender
Direct care	Age, occupation, gender
Supervisor support	Age, direct care, pre-COVID mental health, occupation, gender
Community infection risk	Age, COVID diagnosis, pre-COVID mental health, physical health, gender
Work infection risk	Age, community infection risk, COVID diagnosis, direct care, pre-COVID mental health, occupation, physical health, gender, supervisor support
COVID-19 infection worries	Age, community infection risk, COVID diagnosis, direct care, pre-COVID mental health, occupation, physical health, gender, supervisor support, work infection risk
COVID diagnosis	Age, direct care, gender, supervisor support
Outdoors time	COVID infection worries, age, community infection risk, COVID diagnosis, pre-COVID mental health, physical health, gender, work infection risk
Stress from lockdown restrictions	COVID infection worries, age, community infection risk, COVID diagnosis, pre-COVID mental health, occupation, physical health, changes in relationship quality, gender, supervisor support, work infection risk
Relationship stress	COVID infection worries, age, community infection risk, COVID diagnosis, pre-COVID mental health, outdoors time, physical health, changes in relationship quality, gender, stress from lockdown restrictions, supervisor support, work infection risk
Changes in relationship quality	COVID infection worries, age, community infection risk, COVID diagnosis, pre-COVID mental health, outdoors time, physical health, gender, supervisor support, work infection risk

*Minimally sufficient adjustment set: the minimum number of covariates needed to identify the total effect without inducing biasing associations.