Table 2.
Mixed linear regression models with random effects at grantee level predicting change at one month in PHQ-9 score (shown on the left) and GAD-7 score (shown on the right)
| Variable | PHQ-9 Score Changea | GAD-7 Score Changea | ||||
|---|---|---|---|---|---|---|
| Coefficient | P -value | 95% CI | Coefficient | P -value | 95% CI | |
| Telehealth groupb | 0.05 | -0.53, 0.62 | 0.19 | -0.35, 0.74 | ||
| Baseline score | -0.38 | 0.001 | -0.42, -0.34 | -0.35 | 0.001 | -0.40, -0.31 |
| Age group (0–18 as reference) | ||||||
| Age group 19–34 | -0.58 | -1.54, 0.38 | -0.23 | -1.08, 0.62 | ||
| Age group 35–64 | -0.77 | -1.70, 0.16 | -0.31 | -1.13, 0.50 | ||
| Age group 65+ | -1.48 | 0.05 | -2.70, -0.27 | -0.95 | -2.03, 0.13 | |
| Sex (Male as reference) | ||||||
| Female | 0.42 | -0.15, 0.99 | 0.80 | 0.01 | 0.23, 1.36 | |
| Race (White as reference) | ||||||
| Non-Whites | 0.30 | -0.74, 1.35 | 0.35 | -0.65, 1.35 | ||
| Unknown race | -0.55 | -1.59, 0.49 | -0.34 | -1.31, 0.63 | ||
| Ethnicity (Hispanic as reference) | ||||||
| Non-Hispanic | -0.18 | -1.46, 1.10 | -0.02 | -1.23, 1.18 | ||
| Unknown ethnicity | -0.39 | -1.93, 1.15 | -0.24 | -1.65, 1.17 | ||
| Insurance status (Private ins. as reference) | ||||||
| Public insurance | 0.29 | -0.36, 0.95 | 0.37 | -0.22, 0.96 | ||
| Self-pay/unknown /other insurance | 0.19 | -0.76, 1.15 | 0.01 | -0.82, 0.80 | ||
| N | 1,514 | 1,284 | ||||
aBoth models adjust for age, gender, race, ethnicity, insurance type, and baseline score. Additionally, random effects are specified at the grantee level. None of the covariates are significantly related to the outcome variable (last follow-up change scores) other than baseline scores in both models and self-pay/unknown/other insurance in the subsample model
bTelehealth treatment group is a binary variable coded as 1 for patients in the telehealth treatment group and as 0 for patients in the in-person control group