Individual item differences between these groups are presented in Additional file 2. To interrogate the DREEM data, Miles et al. [ 17] suggest the results are investigated at three levels: i) overall; ii) subscales; iii) items.
For performance related to the third item, differences between baseline and test performances were insignificant for both males (t (5) = 0.23, P = 0.826) and females (t (5) = 2.37, P = 0.063).
Differential item functioning (DIF) analyses were performed, which compared the item differences between two groups of individuals according to socio-demographic parameters (gender, age, education level, disease duration, and living arrangement) to check whether all items behave the same way [ 31].
A Principal Component Analysis (PCA) of the residuals is undertaken, the standardised person-item differences between the observed data and what is expected by the model for every person's response to every item.
In order to test the main hypotheses, for total ISAHCQ scores as well as for single items, differences between pre-, post-, and follow-up measurements were analysed by a non-parametric test for paired ordinal data (Wilcoxon signed-rank test).
To perform tests of significance on responses to matched items, we calculated the standardized residuals as post hoc analyses on matched items (the difference between the observed and expected values converted to a z score).
Although interim testing led to better performance across all items, the difference between the two groups was largest for neutral and misleading trials (consistent: t 86) = 2.58, P = 0.01 (ns after Bonferroni correction); neutral: t 86) = 6.67, P < 0.001, d = 1.44; misleading: t 86) = 4.43, P < 0.001, d = 0.90).
Furthermore, no significant item-memory difference between patients and controls was seen: t(29) = 1.25, p = 0.22.
Item discrimination is the measure of how an item detects differences between higher and lower scores on a test.
The study results are context-specific due to the intervention's resource utilisation being dependent on e.g. infrastructure, professional roles, participation rates and item cost differences between markets.
