Skip to main content

Table 2 Long-term effect on self-rated health from unemployment at 40 years of age for different estimators and unemployment mode

From: Methodological perspectives on the study of the health effects of unemployment – reviewing the mode of unemployment, the statistical analysis method and the role of confounding factors

Unemployment

mode

Unemployment during

follow-upb

 

Estimatora

 

Employed / unemployed

Logistic regression

G-computation

IPW standard

IPW augmented

IPW

doubly robust

MSEc,

low/high

All confoundersd

Self-reported long-term unemployment

No censor

627 / 178

1.90

(1.31–2.76)

0.134

(0.051–0.22)

0.129

(0.045–0.21)

0.128

(0.042–0.21)

0.125

(0.041–0.21)

0.0018/ 0.0019

Censor

522 / 98

1.73

(1.07–2.81)

0.111

(0.008–0.22)

0.114

(0.001–0.25)

0.108

(-0.005–0.23)

0.124

(0.022–0.23)

0.0028/ 0.0038

Register-based long-term unemployment

No censor

585 / 214

1.65

(1.16–2.34)

0.103

(0.027–0.18)

0.104

(0.028–0.18)

0.102

(0.025–0.18)

0.099

(0.025–0.17)

0.0014/ 0.0015

Censor

488 / 122

1.38

(0.88–2.17)

0.064

(-0.027–0.16)

0.067

(-0.025–0.17)

0.069

(-0.025–0.16)

0.074

(-0.015–0.17)

0.0022/ 0.0023

Current unemployment

No censor

693 /61

1.76

(0.97–3.18)

0.121

(-0.018–0.26)

0.199

(0.006–0.48)

0.253

(0.057–0.44)

0.252

(-0.003–0.43)

0.0052/

0.0147

Significant confounderse

Self-reported long-term unemployment

No censor

627 / 178

1.85

(1.28–2.67)

0.129

(0.049–0.21)

0.123

(0.042–0.20)

0.123

(0.041–0.20)

0.120

(0.040–0.20)

0.0018/ 0.0021

Censor

522 / 98

1.74

(1.08–2.79)

0.113

(0.013–0.22)

0.103

(0.0005–0.21)

0.103

(0.0002–0.21)

0.101

(0.004–0.20)

0.0025/ 0.0027

Register-based long-term unemployment

No censor

585 / 214

1.63

(1.16–2.31)

0.102

(0.028–0.17)

0.096

(0.022–0.17)

0.097

(0.022–0.17)

0.095

(0.021–0.17)

0.0014/ 0.0014

Censor

488 / 122

1.40

(0.89–2.16)

0.066

(-0.022–0.16)

0.059

(-0.029–0.15)

0.059

(-0.029–0.15)

0.057

(-0.028–0.15)

0.0020/ 0.0021

Current unemployment

No censor

693 /61

1.66

(0.94–2.91)

0.109

(-0.019–0.25)

0.156

(-0.003–0.33)

0.155

(-0.005–0.32)

0.163

(0.012–0.32)

0.0046 /0.0072

  1. IPW = inverse probability weighting estimator; Risk differences are used for the G-computation and IPW estimators
  2. The p-value is below 0.05 if 1 is not in the confidence interval for logistic regression estimator and 0 not in the confidence interval for other estimators. All variables but the outcome variable was measured in 1995
  3. a Confidence intervals, corresponding to the 2.5% and 97.5 Bootstrap percentiles, are presented within parentheses. Estimates represent the health effect on unemployed compared to employed individuals as represented with relative risk for logistic regression (above 0 a poorer health) and risk difference (above 1 a poorer health) for other estimators
  4. b Participants were censored or not censored if they experienced unemployment between 1995 and 2007
  5. c MSE = Mean squared error. The lowest and highest MSE for the estimators of the definition is presented
  6. d Analyzes are controlling for alcohol intake, Availability of Attachment (AVAT), Availability of Social Integration (AVSI), body mass index, cash margin, education level, gender, marital status, previous health status, occupation and smoking
  7. e Significant confounders were determined with logistic regression. Analyzes are controlling for education level, marital status, previous health status and occupation
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us