Response bias influences mental health symptom reporting in patients with obstructive sleep apnea

Annals of Behavioral Medicine (Impact Factor: 4.2). 11/2001; 23(4):313-317. DOI: 10.1207/S15324796ABM2304_11

ABSTRACT The Medical Outcomes Study (MOS) inventory is widely used to assess quality of life in chronically ill patients. Although
response bias was addressed during its initial validation, we are unaware of subsequent studies examining how personality
characteristics influence responses on this instrument. We examined the impact of response bias on MOS data reported by 44
obstructive sleep apnea (OSA) outpatients who completed the MOS and Marlowe—Crowne Social Desirability Scale (MC). Social
desirability implies responding in a way that the participant believes will be viewed favorably; MC is often used to statistically
control for this form of response bias on self-report measures. The MOS yielded 2 summary scales (Mental and Physical Health)
and 8 dimensions (based on Short Form-36 [SF-36]). Data were analyzed using multiple regression analyses. Results show that
there is a significant relation between MC and the Mental and Physical Health factors (R2 = .194, F = 5.069, p = .011) and the eight SF-36 dimensions (R2= .359, F =2.448, p = .032). Post hoc analyses did not reveal that any one independent variable was a superior predictor of
the MC. Participants scoring high on the MC reported 31% greater health than participants scoring low on the MC. Findings
suggest that response bias has a significant influence on MOS data from OSA patients, with the greatest impact on mental health
indexes. Controlling for response bias on the MOS and other self-report measures is important in both research and clinical
situations with OSA and potentially other chronically ill patients.

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Objective. To assess the role of personality as a predictor of Short form-36 (SF-36) in distressed patients (perceived stress questionnaire, PSQ) with ulcerative colitis (UC) and Crohn's disease (CD). Material and methods. Fifty-four patients with CD and 55 with UC (age 18–60 years) who had relapsed in the previous 18 months, i.e. with an activity index (AI) for UC or CD≥4, PSQ≥60, and without severe mental or other major medical conditions, completed the Buss-Perry Aggression Questionnaire (BPA), the Neuroticism and Lie scales of the Eysenck Personality Questionnaire (EPQ-N and -L), the Multidimensional Health Locus of Control Scale (LOC) (Internal (I), Powerful Other (PO), Chance (C)), the Toronto Alexithymia Scale (TAS) and the SF-36. Results. Multiple linear regression analyses controlling for gender, age and clinical disease activity (AI) in separate analyses for UC and CD showed that the mental and vitality subscales were predicted by neuroticism in both UC and CD. The highest explained variance was 43.8% on the “mental” subscale in UC. The social function subscale was related to alexithymia only in UC, while the role limitation and pain subscales were related to personality in CD only. The physical function subscale related differently to personality in UC and CD. Conclusions. While mental and vitality subscales were predicted by neuroticism in both UC and CD, other subscales had different relationships to personality, suggesting different psychobiological interactions in UC and CD.
    Scandinavian Journal of Gastroenterology 07/2009; 43(12):1505-1513. DOI:10.1080/00365520802321196 · 2.33 Impact Factor
  • Source
  • [Show abstract] [Hide abstract]
    ABSTRACT: Evidence suggests substantial comorbidity between symptoms of somatization and depression in clinical as well as nonclinical populations. However, as most existing research has been retrospective or cross-sectional in design, very little is known about the specific nature of this relationship. In particular, it is unclear whether somatic complaints may heighten the risk for the subsequent development of depressive symptoms. We report findings on the link between symptoms of somatization (assessed using the SCL-90-R) and depression 5 years later (assessed using the CES-D) in an initially healthy cohort of community adults, based on prospective data from the RENO Diet-Heart Study. Gender-stratified multiple regression analyses revealed that baseline CES-D scores were the best predictors of subsequent depressive symptoms for men and women. Baseline scores on the SCL-90-R somatization subscale significantly predicted subsequent self-reported symptoms of depressed mood 5 years later, but only in women. However, somatic complaints were a somewhat less powerful predictor than income and age. Our findings suggest that somatic complaints may represent one, but not necessarily the most important, risk factor for the subsequent development of depressive symptoms in women in nonclinical populations. The results also highlight the importance of including social variables in studies on women's depression as well as conducting additional research to further examine predictors of depressive symptoms in men.
    Psychotherapy and Psychosomatics 01/2003; 72(5):261-7. DOI:10.1159/000071897 · 9.37 Impact Factor