Table 2

Pathways between osteoarthritis and mortality via walking frequency, depression, anxiety and unrefreshed sleep: total, direct and indirect effects

MediatorAssociation between osteoarthritis and potential mediator
Adjusted OR (95% CI)PathwayAdjusted HR (95% CI)
Total effect (no mediators)1.11(1.06to1.17)
Walking frequencyDirect1.05(1.00to1.10)
 Frequent walkerReferenceIndirect1.05(1.04to1.06)
 Non-frequent walker1.89(1.82 to1.97)*Total1.10(1.05to1.15)
DepressionDirect1.08(1.03to1.13)
 No depressionReferenceIndirect1.02(1.02to1.03)
 Depression2.10(2.00 to2.21)Total1.11(1.06to1.16)
AnxietyDirect1.10(1.05to1.15)
 No anxietyReferenceIndirect1.01(1.00to1.02)
 Anxiety1.96(1.88 to2.05)Total1.11(1.06to1.16)
Unrefreshed sleepDirect1.09(1.05to1.14)
 Refreshed sleepReferenceIndirect1.01(1.00to1.01)
 Unrefreshed sleep1.98(1.88 to2.09)§Total1.12(1.08to1.17)
  • All models adjusted for age, sex, education, occupation, smoking status, cognitive behaviour, BMI, consultation for ischaemic heart disease, chronic obstructive pulmonary disease, diabetes, non-steroidal anti-inflammatories and self-report of heart conditions.

  • Significant associations between predictor and mediator and indirect effects are in bold.

  • *E value for point estimate=2.09; E value for lower CI=2.04.

  • †E value for point estimate=2.26; E value for lower CI=2.18.

  • ‡E value for point estimate=2.15; E value for lower CI=2.08.

  • §E value for point estimate=2.16; E value for lower CI=2.08.

  • BMI, body mass index.