Participants Characteristics and Descriptive Results
Descriptive analyses (Table 1) revealed a significant difference in the average ages, and sex distributions of both groups and showed that MD patients had also started to learn their instruments significantly later.
Table 1
Descriptive characteristics and questionnaire scores of both groups
Parameter
|
MD (n=128)
|
HM (n=136)
|
|
|
Mdn (IQR)
|
Mdn (IQR)
|
Test statistic
|
p-value
|
Age in years (M±SD)
|
56.13 (10.63)
|
38.03 (12.27)
|
14795
|
.000***
|
Sex: male/female/other (n)
|
92/36/0
|
55/81/0
|
χ2(1, N=264) = 25.14
|
.000***
|
Age started playing the instrument
|
9.5 (5)
|
7.0 (4)
|
12314
|
.000***
|
Current occupation: professional/student/other/NA (n)
|
117/2/5/4
|
106/30/0/0
|
|
|
Dystonia onset: age (M±SD)
|
38.46 (11.73)
|
-
|
|
|
Subjective motor disability at instrument (M±SD)
|
3.66 (1.19)
|
-
|
|
|
Subjective motor disability in everyday life (M±SD)
|
1.57 (0.90)
|
-
|
|
|
CTQ
|
|
|
|
|
|
Total Score
|
33.50 (12.25)
|
32 (8)
|
10292
|
.052
|
|
Emotional abuse
|
6 (4)
|
6.5 (3)
|
8834.5
|
.830
|
|
Physical abuse
|
5 (1)
|
5 (0)
|
9730
|
.110
|
|
Sexual abuse
|
5 (0)
|
5 (0)
|
9521
|
.110
|
|
Emotional neglect
|
10 (6)
|
8 (5)
|
11046
|
.000***
|
|
Physical neglect
|
6 (3)
|
5 (2)
|
9740.5
|
.154
|
ACE score
|
0 (0)
|
0 (0)
|
8849
|
.799
|
FMPS
|
|
|
|
|
|
Total Score (without O)
|
78 (34.25)
|
82 (28)
|
8336.5
|
.554
|
|
Concerns about Mistakes and Doubts
|
38 (19)
|
38 (16.25)
|
8525
|
.773
|
|
Parental Expectations and
Criticism
|
20 (12.25)
|
20 (12)
|
8489
|
.729
|
|
Personal Standards
|
23.5 (10)
|
24 (8)
|
8271
|
.485
|
|
Organization
|
23 (6)
|
24 (5)
|
8200.5
|
.415
|
Note. Subjective motor disability measured on a 5-point Likert-scale ranging from 1 (no/minimal disability) to 5 (maximum disability/unable to play). Median and interquartile ranges are displayed, unless otherwise indicated. Test statistics show W of the Wilcoxon rank-sum test unless otherwise indicated. p-values for CTQ and FMPS scales were Holm-corrected. ***p < .001.
Abbreviations: MD = musician’s dystonia patients, HM = healthy musicians, CTQ = Childhood Trauma Questionnaire, ACE = Adverse Childhood Experiences, FMPS = Frost’s Multidimensional Perfectionism Scale, O = Organization.
|
Descriptive analysis further revealed higher rates of emotional neglect and a slightly higher CTQ total score for MD patients. The average emotional neglect score of the MD group can be classified as slight to moderate (classification by Häuser et al.38), while it was none to minimal among the healthy controls. The Wilcoxon rank-sum test showed only the difference on the sub-score emotional neglect (EN) to be significant after Holm correction, the effect for the difference being small (r = .24, 95% CI [0.11, 0.35]). Therefore, apart from the CTQ total score, the sub-score “emotional neglect” was specially observed in further analyses. No significant differences were observed between the median of the ACE scores of both groups (Table 1), but descriptive analysis showed more frequent occurrences of almost all forms of ACEs among the MD group (Figure 1).
Especially “emotional abuse” was reported significantly more often by the MD group (χ2(1, N = 264) = 4.5, p = .03). The item ACE11 (“other stressful experiences”) was reported by 37.5% (n=48) of the MD patients and by 26.5% (n=36) of the healthy musicians. Using qualitative thematic analysis, 15 themes were identified from the musician’s written answers to ACE11 (Table 2). Most prominent among the MD group were reports of emotional distance or neglect (11 counts), conflicts among family members that were partly violent (12 counts), emotional abuse from family members (6 counts), and performance pressure at the instrument (6 counts). Themes most often reported by the HM group were bullying at school (8 counts) and non-violent conflicts among family members (7 counts).
Table 2
Qualitative thematic analysis of "other stressful experiences"
Identified themes
|
Absolute numbers of reports (%)
|
HM (n=136)
|
MD (n=128)
|
Emotional distance/neglect
|
2 (1,5)
|
11 (8.6)
|
Emotional and psychological abuse
|
1 (0.7)
|
6 (4.9)
|
Conflicts among family members
|
7 (0.7)
|
7 (5.5)
|
Victim or witness of physical violence
|
2 (1,5)
|
5 (3.9)
|
Separation from parent(s)/
Death of a parent, family member or close friend
|
3 (2.2)
|
6 (4.9)
|
(Assumed) sexual abuse
|
1 (0.7)
|
4 (3.1)
|
Mental illness in childhood
|
3 (2.2)
|
6 (4.9)
|
Physical illness in childhood
|
-
|
2 (1.6)
|
Mental disorder of a parent
|
1 (0.7)
|
5 (3.9)
|
Alcohol abuse of a parent
|
1 (0.7)
|
5 (3.9)
|
Divorce/separation of parents
|
2 (1,5)
|
2 (1.6)
|
Bullying at school/in the neighbourhood
|
8 (5.9)
|
2 (1.6)
|
Hospital stay alone
|
2 (1,5)
|
1 (0.8)
|
Performance pressure, negative experiences at the instrument
|
4 (2.9)
|
6 (4.9)
|
Other
|
2 (1,5)
|
3 (2.5)
|
Total number of reported events
|
39
|
71
|
Note. Themes were identified according to the written answers participants gave to the question “Were there any other events in your childhood or adolescence that you would rate as extremely stressful?” |
There were no significant differences between MD and HM in the total score of Frost’s Multidimensional Perfectionism Scale and its sub-scores.
Regarding subjective motor disability at the instrument, descriptive analysis revealed slightly higher CTQ and emotional neglect scores for those participants with a subjectively more severe motor disability, but patients who rated their disability to be moderate (3) on average scored the highest total score on the CTQ (Mdn(IQR) = 40(10.25)) and on the sub-score emotional neglect (Mdn(IQR) = 13(5.25)) (Figure 2).
Spearman’s rank correlation revealed the CTQ total score and EN to be amongst others positively correlated with age, disorder, subjective disability and the FMPS score (Supplementary Table S1).
Logistic Regression
Linear regression revealed significant associations between CTQ total score and age (β(SE) = .10 (.04), t(260) = 2.46, p < .05), as well as CTQ total score and age at the start of playing ((β(SE) = .43 (.17), t(260) = 2.46, p < .05). The association between age and CTQ total score, however, was only observable for the total sample. When both groups (MD and HM) were analyzed separately, the CTQ total score was not associated with the age of the participants, which gives additional support to the decision not to include age as covariate, as described in the methods section.
The participant’s sex was not found to be associated with ACEs.
In the restricted logistic regression model with only the CTQ total score as predictor (Model 1a), the CTQ total score was found to significantly contribute to the prediction of MD and the model was found to be a significant improvement to the baseline model (see Table 3). The explanatory value of the CTQ total score remained significant when sex, and age at the start of playing were additionally included into the model (Model 1b). Compared to a model with only sex and age at the start of playing as predictors, including the CTQ total score into the model contributed significantly to an improvement of its explanatory value (χ2(df) = 7.03(2), p < .01). Model 1b can be viewed as acceptable according to the calculated pseudo R2 = .26.
Table 3
Predictors of musician’s dystonia in uncontrolled and controlled logistic regression models.
Variables
|
Model 1a
|
Model 1b
|
β (SE)
|
OR
|
95% CI for odds ratio
|
β (SE)
|
OR
|
95% CI for odds ratio
|
LL
|
UL
|
LL
|
UL
|
Constant
|
-0.05 (0.13)
|
|
|
|
-0.66 (0.21)**
|
|
|
|
CTQ Score
|
0.05 (0.01)**
|
1.05
|
1.02
|
1.08
|
0.04 (0.02)*
|
1.04
|
1.01
|
1.08
|
Age at start of playing
|
|
|
|
|
0.20 (0.05)***
|
1.23
|
1.12
|
1.35
|
Sex
|
|
|
|
|
1.13 (0.28)***
|
3.10
|
1.80
|
5.43
|
|
|
|
|
|
|
|
|
|
χ2 (df)
|
10.73 (1)**
|
|
|
|
57.95 (3)***
|
|
|
|
Nagelkerke R2
|
.05
|
|
|
|
.26
|
|
|
|
|
Model 2a
|
Model 2b
|
Constant
|
-0.05 (0.13)
|
|
|
|
-0.68 (0.22)**
|
|
|
|
Emotional abuse
|
-0.10 (0.06)
|
0.91
|
0.80
|
1.02
|
-0.02 (0.07)
|
0.98
|
0.86
|
1.12
|
Physical abuse
|
0.12 (0.10)
|
1.12
|
0.92
|
1.39
|
-0.00 (0.12)
|
1.00
|
0.78
|
1.26
|
Sexual abuse
|
0.04 (0.08)
|
1.04
|
0.90
|
1.24
|
0.12 (0.09)
|
1.13
|
0.96
|
1.38
|
Emotional neglect
|
0.17 (0.05)***
|
1.19
|
1.09
|
1.30
|
0.12 (0.05)*
|
1.13
|
1.02
|
1.25
|
Physical neglect
|
-0.04 (0.08)
|
0.96
|
0.82
|
1.11
|
-0.07 (0.08)
|
0.94
|
0.80
|
1.10
|
Age at start of playing
|
|
|
|
|
0.20 (0.05)***
|
1.22
|
1.11
|
1.25
|
Sex
|
|
|
|
|
1.16 (0.30)***
|
3.20
|
1.79
|
5.81
|
|
|
|
|
|
|
|
|
|
χ2 (df)
|
21.32 (5)***
|
|
|
|
62.66 (7)***
|
|
|
|
Nagelkerke R2
|
.10
|
|
|
|
.28
|
|
|
|
Note. OR = odds ratio; CI = confidence interval; LL = lower level; UL = upper level. Model 1a and 2a: crude associations. Model 1b and 2b: adjusted for sex and age at the start of playing. χ2 statistics show relation of explanatory value of the model compared to the baseline model.
***p < .001; **p < .01; *p < .05.
Abbreviations: CTQ = Childhood Trauma Questionnaire.
|
In a closer examination of the CTQ sub-scores (Table 3), emotional neglect was identified as the single significant predictor for musician’s dystonia of all sub-scores in an uncontrolled model (Model 2a), as well as in a model additionally including sex and age at the start of playing. The calculated pseudo R2 =.28 for the controlled model (Model 2b) again showed an acceptable fit of the model.
In a separate regression model, the ACE score was not found to have a significant explanatory value for musician’s dystonia in musicians (β = 0.13, SE = 0.09, Model χ2(df) = 2.12(1), p = .15), and neither was the FMPS score (β = -0.00, SE = 0.01, Model χ2(df) = 0.09(1), p =.76).
Ordinal Regression
In the ordinal logistic regression model uncontrolled for covariates, the CTQ total score was found to be a significant predictor for a higher rating of subjective disability (β = 0.04, SE = 0.01, Wald = 3.14, p < .01) and the same held true for emotional neglect (β = 0.12, SE = 0.03, Wald = 4.25, p < .001). The odds ratio for rating subjective disability one category higher, with a higher CTQ total score of one unit, was 1.04 (95% CI [1.01, 1.06]), while it was 1.13 (95% CI [1.07, 1.19]) for a higher score of one unit on the scale emotional neglect.
However, in an adjusted model including the CTQ total score, age, sex, and age at the start of playing, current age (β = 0.10, SE = 0.01, Wald = 8.12, p < .001) and age at the start of playing (β = 0.09, SE = 0.04, Wald = 2.13, p < .05) were found to be the only predictors contributing to higher ratings of subjective motor disability (for graphic analysis, see Supplementary Figure S1).
The ACE score again had no influence on the probability to experience a certain severity of subjective disability (β = 0.09, SE = 0.08, Wald = 1.12, p = .26). The FMPS score showed to have no influence on subjective motor disability alone (β = 0.00, SE = 0.01, Wald = 0.21, p = .84) and no interaction of the FMPS score with the CTQ total score was found (β = 0.00, SE = 0.01, Wald = 1.61, p = .11).
Mediator Analysis
No mediating effects of the FMPS score between the CTQ total score and musician’s dystonia were observed. While there was a significant association between CTQ total score and FMPS score with the CTQ explaining 14% of the variance in perfectionism (R2 = .14, F(1,262) = 44.26, p < .001), the FMPS score had no significant effect on musician’s dystonia (see Figure 3).
The indirect effect from CTQ on musician’s dystonia through perfectionism was -0.00 (95% CI [-0.00, 0.00]), showing no increase of probability. The significance of the indirect effect was tested using a bootstrap procedure39, resampling the indirect effects 1,000 times, which revealed the indirect effect to be not significant (p = .10).