. 2023 Jan 4;6(1):e2249731. doi: 10.1001/jamanetworkopen.2022.49731

Table 2. Association of an Additional Social Worker With Individuals per 1000 Veterans Who Received Palliative Care or Hospice Services in the Month After a Hospital Stay by Type of Hospitalization, October 2016 to September 2019.

Model	Change, No. per 1000 veterans (95% CI)
Model	Any hospitalization	VA hospitalization	Community hospitalization
2-Way fixed-effects model^a	5.3 (2.3 to 8.3)	5.3 (1.4 to 9.1)	6.0 (−0.2 to 12.1)
P value	.001	.009	.06
Double-robust (group-time average)^b	15.6 (9.2 to 22)	14.3 (5.9 to 22.7)	10.4 (4.1 to 16.8)

^{^a}

Adjusted for site and year-month fixed effects and demographic and clinical characteristics listed in Table 1.

^{^b}

Change in outcome for intervention sites is compared with not-yet-treated sites. Association is estimated separately for each group of sites that received the intervention in the same month, and these estimates are combined in a weighted mean based on the size of each group.