Table 1.
Socio-demographic and clinical characteristics of the study cohort, by rurality of residence.
Table 2.
Unadjusted percentage of Veterans in the study cohort with at least 1 VA visit for indicated reasons, pre- and post-universal screening implementation.
Table 3.
Patterns of screening or evaluation by for Veterans who had at least one mental health visit in the post-universal screening implementation period.
Fig 1.
Unadjusted trends in the monthly likelihood of a suicide screen.
Fig 2.
Model-adjusted outcome differences compared to Nov. 2020 –monthly coefficients from regression analyses.
Note: Analyses not conditioned on positive C-SSRSs included 1,654,180 Veterans and 33,083,600 Veteran-monthly observations. Analyses conditioned on positive C-SSRSs included 57,673 Veterans and 85,237 Veteran-monthly observations. All models adjusted for Veterans’ age, sex, race, ethnicity, rurality of residence, number of physical and mental health chronic conditions, diagnoses of substance use disorder, post-traumatic stress disorder and depression, Nosos score, VA priority-based enrollment, marital status, homelessness indicator, high suicide risk indicator, cumulative monthly COVID-19 cases in the patient’s county. All models included indicators for patients’ closest facility to control for any time-invariant facility characteristics. In sensitivity analyses, models also adjusted for broadband coverage in patients’ residential zip-codes.
Fig 3.
Model-adjusted outcome differences compared to Nov. 2020, for rural and urban Veterans–monthly coefficients from regression analyses.
Note: Rural (urban) cohort analyses not conditioned on positive C-SSRSs included 485,592 (1,168,588) Veterans and 9,711,840 (23,371,760) Veteran-monthly observations. Rural (urban) cohort conditioned on positive C-SSRSs included 15,097 (42,576) Veterans and 21,040 (64,197) Veteran-monthly observations. All models adjusted for Veterans’ age, sex, race, ethnicity, number of physical and mental health chronic conditions, diagnoses of substance use disorder, post-traumatic stress disorder and depression, nosos score, VA priority-based enrollment, marital status, homelessness indicator, high suicide risk indicator, cumulative monthly COVID-19 cases in the patient’s county. All models included indicators for patients’ closest facility to control for any time-invariant facility characteristics. In sensitivity analyses, models also adjusted for broadband coverage in patients’ residential zip-codes.