Skip to main content
. 2014 Oct 14;111(11):2152–2155. doi: 10.1038/bjc.2014.507

Table 1. Multivariable multilevel logistic regression of factors associated with adherence to repeat testing in the CRC screening programme (i.e., ‘compliant participants' vs ‘occasional participants', Haut-Rhin, France 2007–2010, n=61 386).

   
Full multilevel model
  Empty model ORa (95% CI) P-valuea
Fixed effects
Level 1: individual      
 Gender     0.0001
  Women   1.00  
  Men
 
0.90 (0.86–0.94)
 
 Age (years)     <0.0001
  50–54   1.00  
  55–59   1.32 (1.23–1.43)  
  60–64   1.58 (1.46–1.71)  
  65–69   1.75 (1.61–1.90)  
  70–74
 
1.95 (1.79–2.13)
 
 Health insurance plan     <0.0001
  General plan   1.00  
  Self-employed plan   0.85 (0.76–0.95)  
  Agricultural plan   1.05 (0.95–1.17)  
  Special plan   1.43 (1.27–1.60)  
  Civil service plan
 
1.21 (1.13–1.30)
 
 Source of invitation     <0.0001
  Mail   1.00  
  General practitioner
 
10.73 (10.01–11.50)
 
Level 2: area      
 EDI (deprivation index)     <0.0001
  Quintile 1 (least deprived)   1.00  
  Quintile 2   1.05 (0.98–1.12)  
  Quintile 3   1.00 (0.93–1.08)  
  Quintile 4   1.04 (0.96–1.13)  
  Quintile 5 (most deprived)
 
0.75 (0.70–0.80)
 
 Rural/urban areas     0.0409
  Rural   1.00  
  Urban
 
0.94 (0.88–1.00)
 
Random effects
Level 2 variance (s.e.) 0.2709 (0.0559) 0.1133 (0.0366)

Abbreviations: CRC=colorectal cancer; ORa=odds ratio adjusted for all the parameters listed in the table; 95% CI=95% confidence interval; EDI=European Deprivation Index

a

P, P-trend for ordinal variables; P of heterogeneity for nonordinal variables.