Multiple Sclerosis Journal - October 2017 - 1530

Multiple Sclerosis Journal 23(11)
Other parameters. In the ECM group only, the
Energy Conservation Strategies Survey (ECSS)26 was
administered once, directly after the intervention. It
assesses participants' use of 14 suggested energy conservation strategies that were emphasized in the ECM
course. It rates strategy use as a direct result of the
course and the perceived effectiveness (scale of
1 = not effective to 10 = very effective).
Exacerbations during the study were registered if confirmed by a neurologist and treated with medication.
Serious adverse events (SAEs) were registered by the
principal investigator and were reported to the
Medical Ethical Committee. The definition of an SAE
is included in the design paper.16
Statistical analysis
Statistical analysis was performed by a researcher
blinded to the group allocation. Effectiveness of the
ECM intervention was analysed on a modified intention-to-treat (ITT) principle, using data on each randomized subject with at least one post-randomization
measurement. We used linear mixed models with a
three-level structure (repeated measures, patients and
therapists). First, we evaluated the between-group differences on average during the 1-year study with
group allocation and baseline values of the particular
outcome variable as covariates (overall intervention
effect). Then, we assessed the between-group differences at the separate four follow-up measurements
(T8, T16 and T26 and T52) by adding time (treated as
a categorical variable and represented by dummy variables), and an interaction between group allocation
and time to evaluate the effect of ECM at the specific
time points (time-specific intervention effects). With
the time × group interaction analyses, the differences
between the groups in changes within all time intervals were evaluated. As we used mixed model analyses, no imputation of missing data was performed.27
Additionally, we used linear mixed models to analyse
the within-group changes of the two primary outcomes (CIS20r and IPA) between baseline and T16
and between baseline and T52 for both the ECM
group and the control group.
We present the difference between groups (B),
p-values and 95% confidence intervals (CIs) for the
crude models and for models adjusted for centre,
gender, exacerbations during the study (yes or no)
and time since diagnosis (in years). IBM SPSS
Statistics version 22 (Chicago, IL) was used for statistical analysis. A p-value of ⩽0.05 was considered
significant.

1530

Results
Eligible participants were recruited between
November 2011 and March 2014, and 86 fatigued
persons with MS were randomized. Figure 1 shows
the flow diagram. Baseline personal and clinical
characteristics of all participants are presented in
Table 2. In all, 10 persons, 6 in the intervention group
and 4 in the control group, dropped out before the
second measurement (T8). As these individuals
missed all follow-up measurements, we excluded
them from further analysis, thus data from 76 patients
were available for the so-called modified ITT analysis.28 Except for fewer years with MS (mean difference 5 years, p = 0.04), the drop-outs did not differ
significantly (see Table 3). The use and perceived
effectiveness of energy conservation strategies was
administered in 34 participants, 85% of the participants had implemented ⩾6 strategies, with a mean of
10 strategies and they perceived the strategies as
effective (mean rate of 7). During the treatment
period, one serious adverse event (relapse) was
reported in the ECM group and one (ischaemic bone
disease) in the control group. During the follow-up
period, three SAEs were reported in both groups. The
events were reported to and judged by the Medical
Ethics Committee to be not directly associated with
the intervention.
The observed data over time are presented in Table
3 and the primary outcomes are graphically presented in Figure 2. The results of the corresponding
linear mixed model analyses are presented in Table
4. No significant overall or time-specific intervention effects were found for the CIS20r fatigue subscale. The interval-specific time effects were also
non-significant, with the exception of the interval
T8-T26 which showed a larger decrease in the ECM
group (B = −6.2; 95% CI: −10.5, −1.8; see also
Figure 2).
The IPA (all domains) showed no overall or time-specific intervention effects. The only positive overall
effect was for the IPA domain social relations, in
favour of the control group (B = 0.19; 95% CI: 0.03,
0.35), an outcome also reflected in time-specific
effects at T8 and T26. No interval-specific effects
were found for the IPA outcomes.
The results on the secondary fatigue outcomes were
identical to CIS20r fatigue. For the other secondary
outcome measures, no overall or time-specific (see
Tables 5 and 6) intervention effects were noted. In
general, the adjusted analyses did not affect results
(see Tables 4-6).

journals.sagepub.com/home/msj

https://journals.sagepub.com/home/msj

Table of Contents for the Digital Edition of Multiple Sclerosis Journal - October 2017