Follow-up regression analysis incorporating the interaction term (metacognition × misinterpretation) showed that the term explained additional variance in health anxiety.
Follow-up regression analysis will then be used to determine the multivariate predictors of these differences including GCPS, care sector, education, and other sociodemographic variables.
Alleles of the SNPs in both LPAR1 and ANKS1A were associated with more than one autoantibody in the initial regressions, and follow-up regression analysis supported a best-fit model in which these alleles were associated with the presence of the two autoantibodies simultaneously.
Regimen-stratified follow-up regression analysis and examination of regression plots (Fig. 1) indicated that hyperglycemia was positively associated with subsequent depressive symptoms among patients who were prescribed insulin (β = 0.31, P = 0.002, indicating "moderate" effect magnitude) but not among those on oral hypoglycemic medication alone (β = −0.10, P = 0.210).
To identify baseline factors that were independently associated with incident fatty liver at follow up, logistic regression analysis was undertaken, and the odds ratio (OR) and 95% confidence intervals for incident fatty liver were calculated for continuous variables.
At on average eleven year follow up, logistic regression analysis showed a significant association between incontinence and decline in activities of daily living (ADL) (OR =2.37, 95% CI =1.01-5.58) (P=0.04).
Further to test our exploratory hypotheses and refine our conceptual model, increasing from 64 (2×32) to 70 participants at follow up will allow logistic regression analysis to study relationships between baseline predictors and disclosure using 6 variables (plus the group variable), using the minimum of 10 per variable.
In addition to cross-sectional analysis to compare the control and intervention group at the second and last follow up, longitudinal analysis will use multilevel regression and generalized estimation equating (GEE) models with the outcome variables as outlined above.
A follow-up hierarchical regression analysis was performed to investigate whether the correlations differed significantly from each other.
A follow-up multiple regression analysis revealed that the composite WM score did not significantly predict SWAN ratings of hyperactivity/impulsivity.
However, we did examine associations with work performance and absence at follow up and end of treatment within the cohort using multivariate regression analysis.
