Nota van inspraak

First, raw scores on the questionnaires for the groups are presented with descriptive statistics. To test whether the LFN1, LFN2, and CG show differences in functioning, nonparametric Kruskal–Wallis tests were utilized for overall comparisons and Mann–Whitney U tests were utilized for pairwise comparisons. In order to control for alpha error growth in multiple testing, a strict significance level of p < 0.01 commonly used in psychological research was applied. This reduces the chance for a Type I error (leading to a 1% chance of obtaining a false positive result), while limiting the risk for a Type II error. Further, interpretations were more based on effect sizes indicating the magnitude of a finding independent from its statistical significance. Nonparametric tests were chosen since the assumptions of normality and of homogeneity of variance were violated for most variables and groups. Normality was tested through skewness and kurtosis values, the Shapiro–Wilk test, and a visual examination of QQ plots and boxplots. Homogeneity of variance was tested by Levene´s test. The magnitudes of the overall effect of group differences were calculated using the effect size measure eta squared and interpreted as small = 0.01, medium = 0.06, and large = 0.14. [73,74].

The magnitude of pairwise group differences was estimated by Cohen´s r and interpreted as small (0.1 < r < 0.3), medium (0.3 ≤ r < 0.5), or large (r ≥ 0.5) [73]. Further, the proportion of individuals with above average symptom reporting on all complaint-related variables (FLei, BRIEF-A, BDI-II, PSQI, FSS, APLN) was computed and compared. Above average symptom reporting was defined as a score above the provided cut-off values proposed by the questionnaire developers for the FSS (≥4) and PSQI (≥5) and when scoring in the categories of ‘moderate’ and ‘severe’ depressive symptom reporting on the BDI-II. For the APLN, above average symptom reporting was considered when scoring

J. Clin. Med. 2024, 13, 935

12 of 31

in the ‘high’ or ‘very high’ categories provided by the test developers, corresponding to a percentile score of 80 or higher. For the remaining questionnaires (FLei and BRIEF-A), above average symptom reporting was defined as a score one standard deviation above the mean, thus a score equal to or above the 84th percentile. The BRIEF-A provides a normative group with percentile scores, from which the percentile cut-off of 84% could be directly applied. The FLei does not provide normative data and the healthy control sample used by the questionnaire developers was markedly smaller than our comparison sample (n = 97), as well as younger (42.5 years), and presented with somewhat less females (59%) [63]. Accordingly, a cut-off score of one standard deviation above the mean scores of the CG of this present study was used [63]. Finally, summarized domain scores were computed for the cognition, depressive symptoms, sleep, fatigue, and stress domains, showing the proportion of individuals with above average scores on at least one of the domain-specific variables. The frequency of individuals with above average symptom reporting in all groups was compared with Chi-Square tests at a significance level of p < 0.01 to control for alpha error growth in multiple testing.

The size of associations between variables was computed using Cramer’s V and interpreted as small (0.1), medium (0.3), and large (0.5) for 1 degree of freedom and as small (0.07), medium (0.21), and large (0.35) for 2 degrees of freedom [73]. Finally, the number of functional domains for which participants would show above average symptom reporting was calculated based on those summarized domain scores. 3. Results For the raw scores, descriptive statistics, group comparisons, and effect sizes on all complaint variables are presented in Table 2. The proportion of individuals with above average symptom reporting, group differences, and effect sizes for all complaint variables are presented in Table 3. Further, Figure 1 depicts the proportion of individuals in each group showing above average symptom reporting on the five functional domains. Finally, descriptive statistics of the coping strategies, group comparisons, and effect sizes are depicted in Table 4. In the following sections, the overall terms LFN2 and CG are used for different subgroups based on the outcome variables. That means that the term LFN2 and CG will refer to the subgroups LFN2-SA and CG-SA for the measures of cognitive functioning and depressive symptoms, and the subgroups LFN2-SB and CG-SB for the measures of sleep, fatigue, stress, and coping.

The proportion of missing values differed between the three groups and between the single questionnaires (see Table S3). With the online forced entry fill-out format, the LFN2 group and CG showed no missing values, except for on the PSQI sleep questionnaire. Proportions of missing values on the PSQI subcomponents ranged from 0 to 5%. No global score could be computed for 6% of the LFN2 group and 3% of the CG. The LFN1 group showed missing values on all questionnaires. The highest proportions were observed on the PSQI subcomponents (0–11%) and the PSQI global score (22%). This high number of missing values on the PSQI is, to a great extent, due to the four open questions regarding the time that participants spent in bed and spent sleeping. If not filled out in the correct format, some subcomponent scores, and consequently the global score, could not be computed. Proportions of missing values on the other questionnaires ranged between 0 and 9%. In total, 50 participants (9.5%) scored above the cut-offs for invalid answering patterns on the three BRIEF-A validity scales, suggesting possible noncredible reports. An overview per group and scale is provided in Table S4.

Since only some groups completed the BRIEF-A (LFN1, LFN2-SA, and CG-SA), and the proportion of individuals with possibly noncredible answer patterns in the LFN2-SB and CG-SB could therefore not be deduced, these 50 individuals were retained in the data set. In order to control for the effect of including those participants in the study, all subsequently described results were also computed with a data set excluding those participants (please refer to Tables S5–S7). No or only very small changes were observed that had no meaningful effect on the outcomes.

J. Clin. Med. 2024, 13, 935

13 of 31

3.1. LFN-Related Cognitive Complaints 3.1.1. Raw Scores The LFN1 group reported the most complaints on all variables, followed by the LFN2 group and the CG. Specifically, the three groups differed overall significantly on all cognitive variables with large effect sizes (Table 2). Only the BRIEF-A Behavioral Regulation showed a medium effect size. Furthermore, all pairwise comparisons were significant. These comparisons reached between the LFN1 group and CG large (FLei Sum and attention score) and medium (all other comparisons) effect sizes. Between the LFN2 group and CG, medium effect sizes were seen on the FLei sum, attention, and memory scales. 3.1.2. Proportion of Individuals with above Average Symptom Reporting The LFN1 group showed the highest proportions of above average symptom reporting (23 to 64%), followed by the LFN2 group (15 to 44%) and the CG (5 to 18%) (Table 3). All overall differences were significant with large (FLei sum, attention, and memory) and medium effect sizes (all executive function scales). Notable significant pairwise differences of a medium effect size were found between the LFN1 group and CG on all FLei scales and between the LFN2 group and CG on the FLei memory scale. Finally, the groups differed J. Clin. Med. 2024, 13, x FOR PEER REVIEW 16 of 31 a significantly on the summarized cognition domain score with a large effect size.

Further, significant pairwise comparison of a medium effect size between the LFN1 group and CG was observed (Table 3, and a visual overview in Figure 1).

Percentage of individuals (%)

100 90 80 70

78

76

75

62

61

57 58

60

79

55

50

40

33

30

30

26

19

20

5

10

1

0 Cognition

Depression

Stress

Fatigue

Sleep

Domains with above average symptom reporting LFN1

LFN2

CG

Figure 1. Proportion of individuals that show above average symptom reporting on each of the five Figure 1. Proportion that show above average symptom reporting on each of the five functional domainsofinindividuals the three groups. functional domains in the three groups.

3.2. LFN-Related Psychological Complaints 3.2.1. Depressive Symptoms Raw Scores In terms of depressive symptoms, the LFN1 group reported the most depressive symptoms, followed by the LFN2 group and the CG (Table 2). The groups differed overall significantly with a large effect size. Also, all pairwise comparisons were significant, with one medium effect size between the LFN1 group and CG. Proportion of Individuals with Above Average Symptom Reporting A moderate depressive symptomatology was observed for 15% of the LFN1 group, 5% of the LFN2 group, and 1% of the CG (Table 3). The overall group difference was significant, with a medium effect size. Pairwise comparisons were also significant, but with small effect sizes. A severe depressive symptomatology was observed for 4% of the LFN1 group, 1% of the LFN2 group, and 0% of the CG. The only significant difference was observed for the pairwise comparison between the LFN1 group and CG, however, with a small effect size. On the summarized depressive symptom domain score, the groups differed significantly overall with a medium effect size (Table 3, Figure 1). Pairwise comparisons were