Individual differences and health in chronic pain: are sex-differences relevant?

Because psychological variables are known to intercorrelate, the goal of this investigation was to compare the unique association between several well-established psychological constructs in pain research and pain-related outcomes. Sex differences are considered because pain is experienced differently across sex groups.

Participants were 456 consecutive chronic pain patients attending a tertiary pain clinic (mean age = 58.4 years, SD = 14.8, 63.6% women). The study design was cross-sectional. Psychological constructs included personality (NEO-Five Factor Inventory), irrational thinking (General Attitudes and Beliefs Scale), and coping (Social Problem Solving Inventory). Outcomes were pain severity and interference (Brief Pain Inventory) and physical, general, and mental health status (Short Form-36). To decide whether the bivariate analyses and the two-block, multivariate linear regressions for each study outcome (block 1 = age, sex, and pain severity; block 2 = psychological variables) should be conducted with the whole sample or split by sex, we first explored whether sex moderated the relationship between psychological variables and outcomes. An alpha level of 0.001 was set to reduce the risk of type I errors due to multiple comparisons.

The moderation analyses indicated no sex differences in the association between psychological variables and study outcomes (all interaction terms p > .05). Thus, further analyses were calculated with the whole sample. Specifically, the bivariate analyses revealed that psychological constructs were intercorrelated in the expected direction and mostly correlated with mental health and overall perceived health status. In the regressions, when controlling for age, sex, and pain severity, psychological factors as a block significantly increased the explained variance of physical functioning (ΔR2 = .037, p < .001), general health (ΔR2 = .138, p < .001), and mental health (ΔR2 = .362, p < .001). However, unique associations were only obtained for mental health and neuroticism (β = - 0.30, p < .001) and a negative problem orientation (β = - 0.26, p < .001).

There is redundancy in the relationship between psychological variables and pain-related outcomes and the strength of this association is highest for mental health status. The association between psychological characteristics and health outcomes was comparable for men and women, which suggests that the same therapeutic targets could be selected in psychological interventions of pain patients irrespective of sex.