Skip to main content

Table 2 Results of hypothesis testing: parameter estimates with confidence intervals (CIs) of the GEE regression analysesa

From: Electronic health record implementation and healthcare workers’ work characteristics and autonomous motivation—a before-and-after study

Step

 

Work characteristics

Autonomous motivation

Job autonomy

Interdependence

 

β

95% CI

β

95% CI

β

95% CI

1

Time

    

− .01

− .07 to .05

 

Age

    

.01

− .01 to .01

 

Intercept

    

6.10**

5.85 to 6.35

2

Time

− .09*

− .17 to − .01

.24**

.14 to .33

  
 

Age

.01+

.00 to .02

− .01*

− .02 to − .01

  
 

Intercept

4.46**

4.04 to 4.88

5.09**

4.74 to 5.44

  

3

Time

    

− .01

− .08 to .05

 

Job autonomy

    

.12**

.07 to .17

 

Interdependence

    

.07**

.02 to .11

 

Age

    

.01

− .01 to .01

 

Intercept

    

5.23**

4.78 to 5.68

4

Time

− .10

− .30 to .09

.06

− .12 to .24

  
 

Nurses

− .02

− .30 to .26

− .71**

− .97 to − .46

  
 

Allied HCPs

− .33*

− .65 to − .01

− .65**

− .94 to − .36

  
 

Administrators

.19

− .16 to .54

− .25

− .58 to .07

  
 

Time × Nurses

.03

− .20 to .26

.20+

− .03 to .43

  
 

Time × Allied HCPs

− .04

− .29 to .22

.21

− .05 to .47

  
 

Time × Administrators

.11

− .15 to .37

.21

− .11 to .53

  
 

Age

.01+

.01 to .02

.01**

− .02 to − .01

  
 

Intercept

4.55**

4.07 to 5.02

5.63**

5.22 to 6.04

  

5

Time

    

.01

− .16 to .18

 

Job autonomy

    

.13**

.08 to .17

 

Interdependence

    

.08**

.03 to .13

 

Nurses

    

.11

− .07 to .29

 

Allied HCPs

    

.10

− .08 to .28

 

Administrators

    

− .08

− .29 to .14

 

Time × Nurses

    

− .04

− .23 to .15

 

Time × Allied HCPs

    

.01

− .19 to .21

 

Time × Administrators

    

− .07

− .28 to .15

 

Age

    

.00

− .01 to .01

 

Intercept

    

5.08**

4.60 to 5.56

  1. The reference category for time was baseline and for profession was physicians. N = 456
  2. aTo assure that performing multiple statistical tests did not lead to false positives, we also analysed the data on our (long) dataset using the much recommended PROCESS macro by Hayes, where time entered to the model as a binary predictor [60]. This resulted in regression coefficients very similar to those presented in Table 2, but it should be noted that the macro was not developed for repeated measures
  3. +p < .10; *p < .05; **p < .01.