TABLE 2—
Logistic Regression Models of Predictors of Jail-Based Mental Health Treatment Engagement: 8 US Midwestern Jails, 2017–2018
| AOR (95% CI) | |
| Model 1: MH referrala | |
| Race/ethnicity: White | 0.77 (0.38, 1.57) |
| Male | 0.96 (0.44, 2.11) |
| Housing insecurity | 0.62 (0.32, 1.20) |
| Nonmetropolitan county | 4.10 (1.57, 10.68) |
| Prior MH treatment or current Rx | 0.71 (0.36, 1.43) |
| Positive K6 score | 0.79 (0.39, 1.58) |
| Substance misuse | 0.94 (0.47, 1.85) |
| Past y jail | 2.03 (1.02, 4.02) |
| Length of stay | 1.00 (0.99, 1.00) |
| Model 2: MH treatmentb | |
| Race/ethnicity: White | 0.72 (0.46, 1.14) |
| Male | 0.91 (0.59, 1.42) |
| Housing insecurity | 0.68 (0.45, 1.01) |
| Metropolitan county | 8.06 (5.12, 12.67) |
| Prior MH treatment or current Rx | 1.32 (0.86, 2.02) |
| Positive K6 score | 1.20 (0.80, 1.81) |
| Substance misuse | 0.72 (0.48, 1.07) |
| Past y jail | 1.78 (1.19, 2.66) |
| Length of stay | 1.01 (1.01, 1.02) |
Note. AOR = adjusted odds ratio; CI = confidence interval; K6 = Kessler-6; MH = mental health; Rx = medications. Sample size was n = 590.
a
χ2(9, n = 590) = 26.711, P = .002; Nagelkerke R2 = 11.2%, predicted cases = 93.1%.
b
χ2(9, n = 590) = 170.135, P < .001; Nagelkerke R2 = 34.3%, predicted cases = 77.1%.