Table 5.
At Baseline (prior 3 months) | All patients (N = 60) | Dartmouth (n = 31) | UAB (n = 29) | ||||||
Mean (SD) | Mean (SD) | Mean (SD) | p | Effect size | |||||
Hospital days/month | 1.44 (3.0) | 1.7 (3.6) | 1.17 (2.2) | 0.72 | 0.18 | ||||
ICU days/month | 0.26 (1.0) | 0.34 (1.3) | 0.16 (0.6) | 0.32 | 0.18 | ||||
ED visits/month | 0.23 (0.4) | 0.26 (0.4) | 0.21 (0.4) | 0.54 | 0.13 | ||||
Resource Use- Difference from Baseline | |||||||||
All patients (N = 34) | Dartmouth (n = 23) | UAB (n = 11) | |||||||
Mean (SE) | p * | Effect size† | Mean (SE) | p * | Effect size† | Mean (SE) | p * | Effect size† | |
Hospital days/month | −0.89 (0.3) | 0.002 | 0.39 | −1.18 (0.4) | 0.006 | 0.57 | −0.61 (0.4) | 0.13 | 0.24 |
Days/month in ICU, Mean | −0.16 (0.1) | 0.06 | 0.29 | −0.12 (0.1) | 0.23 | 0.26 | −0.21 (0.2) | 0.16 | 0.31 |
ED visits/month, Mean | −0.05 (0.1) | 0.33 | 0.17 | 0 (0.1) | 0.94 | 0.05 | −0.09 (0.1) | 0.26 | 0.26 |
Abbreviations: SD standard deviation, SE standard error, Estimates from longitudinal models fitted with negative binomial distributions (log link), adjusted for baseline, PACIC Patient Activation, and religious preference
* p-values from t-test or Fisher’s exact tests, as appropriate
†Effect size: Cohen’s d (Cohen, 1988), or d-equivalent (Rosenthal & Rubin, 2003) small: d ~ 0.2, moderate d ~ 0.5, large d ~ 0.8