Skip to main content

Table 5 Fractional logistic regression models of benefit-finding and resilience (n = 25)

From: End-of-life and bereavement support to families in cancer care: a cross-sectional survey with bereaved family members

Regressors / measures of model fit

Benefit-finding (coping), single item (1–5)

Resilience (BRS-6), mean score (1–5)

 

AMEa

pb

95%-CIc

AMEa

pb

95%-CIc

Full model:

      

Quality of end-of-life care (CANHELP) (mean)

0.933*

0.025

[0.117; 1.749]

0.426*

0.012

[0.093; 0.759]

Time since death to survey completion (months)

0.112

0.468

[-0.191; 0.415]

0.103

0.309

[-0.095; 0.300]

Family member age (years; range 35–90, mean = 61)

0.008

0.748

[-0.043; 0.059]

   

From minimum age (35 years) to under 50 years

   

0.057

0.138

[-0.018; 0.132]

From 50 years to maximum age (90 years)

   

-0.032**

0.004

[-0.053; -0.010]

Family member is the partner / spouse of the deceased (yes)

0.002

0.997

[-1.199; 1.203]

0.309

0.355

[-0.346; 0.964]

Family member is female (vs. male)

-0.084

0.885

[-1.232; 1.063]

-0.073

0.698

[-0.440; 0.294]

BIC-optimized model:

      

Quality of end-of-life care (CANHELP) (mean; range 1–5)

0.976***

0.000

[0.476; 1.477]

0.503***

0.000

[0.279; 0.728]

Likelihood-based pseudo-R2measures of model fit:

      

Full model:

      

Nagelkerke’s R2, Cox & Snell’s R2, McFadden’s R2

0.153

0.112

0.090

0.085

0.062

0.049

BIC-optimized model:

      

Nagelkerke’s R2, Cox & Snell’s R2, McFadden’s R2

0.146

0.107

0.085

0.041

0.030

0.023

  1. Note. Continuous variables in italics. *p < 0.05, **p < 0.01, ***p < 0.001. aAverage marginal effect.bp-value of type one error.cLimits of 95% confidence interval.
  2. CANHELP = “CANadian Health Care EvaLuation Project– Bereavement version”, score range: 1–5, high score = high satisfaction in EoL care.
  3. BRS-6 = “Brief Resilience Scale”, score range: 1–5; high score = high resilience.