Table 4.

Results of exploratory analysis using a non-parametric Wilcox Rank Sum test to uncover group differences/associations based on levels of penetration of the intervention (high penetration is defined as 75% or more by Q4 after implementation)

Variable	High Median	Low Median	p-value
Continuous Variables
Patient demographics
Age of the target population (mean)	63.5	68.2	0.915
Percentage of the target population that is male	45.0	50.0	0.166
Percentage of the target population with a non-English preferred language.	6.5	4.0	1.000
Fidelity ratings (high is a score of 7 or higher out of 10 and low is under 7)
Project lead rating of consistency	9.2	8.5	0.146
Project lead rating of completeness	8.2	7.5	0.297
Project lead rating of content quality	10.0	8.0	0.018
Project lead rating of process quality	8.0	8.0	0.248
Service and client outcomes reported by providers on post-implementation surveys
Mean workload (% Yes)	0.7	0.8	0.832
Mean value (% Yes)	0.9	0.9	0.304
Categorical Variables
Variable	High Percent	Low percent	p-value
Hospital categorization
Hospital Category			1.000
Academic	41.7	33.3
Community	25.0	22.2
Small	33.3	44.4
Hospital Type			0.229
Acute	75.0	66.7
Rehabilitation	0.0	33.3
Implementation process characteristics
Community of Practice participation			0.080
high (attendance at over 75% of meetings)	75.0	66.7
medium (attendance between 25 and 75%)	25.0	0.0
low (attendance at under 25%)	0.0	33.3
Electronic Medical Record integration			1.000
Yes	41.7	44.4
No	58.3	55.6
Implementation as part of broader QI			1.000
Yes	58.3	55.6
No	41.7	44.4
Multidisciplinary team involved in delivery			0.331
Yes	83.3	55.6
No	16.7	44.4

Variables are all aggregated to the hospital-level and include patient demographics, hospital categorization, differentiating characteristics of each hospital’s implementation process, fidelity ratings provided by projject leads during end-of-project interviews, and service and client outcomes reported by providers through post-implementation surveys