Table 5.

Effect of Guided Care on Physicians’ Ratings of Chronic Care Processes: Regression Coefficients and Effect Sizes in Linear Regression Models

VariableICCaCoefficientb (95% CI)Effect Sizec (95% CI)
CI=confidence interval, ICC = intraclass correlation.
aICCs are not shown if the variance within sites is larger than the variance between sites and the values of ICC become artificially low or negative.25
bRegression coefficient of Guided Care in linear regression models of 1-year scores, adjusting for baseline scores and practice ownership.
cCalculated using Hedges’ d as follows:
 Embedded Image.
 where n1 and n2 are sample sizes of 2 comparison groups, and
 Embedded Image
 where n1 and n2 are the numbers of sample size in 2 groups and df is the degrees of freedom used for a corresponding t value in a linear model.
 95% CI = ES− 1:96se to ES+1:96se; where ES stands for effect size and se is the asymptotic standard error for the effect size:
 Embedded Image
d ρ <0.05.
Satisfaction with patient/family communication0.060.59 (0.13 to 1.06)d0.87 (0.21 to 1.54)
Satisfaction with management of chronic care0.250.23 (−0.20 to 0.66)0.37 (−0.28 to 1.01)
Time spent managing chronically ill patients0.12 (−0.35 to 0.60)0.18 (−0.46 to 0.82)
Knowledge of patients’ personal circumstances0.05 (−0.22 to 0.31)0.12 (−0.52 to 0.76)
Knowledge of patients clinical characteristics0.330.32 (0.01 to 0.63)*0.71 (0.05 to 1.37)
Knowledge about all the special visits0.370.07 (−0.37 to 0.52)0.12 (−0.52 to 0.76)
Helped appointment for referral visit0.57 (−0.07 to 1.20)0.61 (−0.04 to 1.26)
Written information to patients’ specialists0.140.54 (−0.20 to 1.29)0.50 (−0.15 to 1.15)
Useful information received from specialists0.200.27 (−0.13 to 0.68)0.46 (−0.19 to 1.10)