Comparison of Italian- and German-speaking patients with chronic pain
Sample 1 Sample 2
Baseline to 1 (T1), 3 (T2) and 6 month (T3) follow-up Baseline to 1 (T1) and 12 month (T4) follow-up
German-speaking patients
Inclusion from January 2001
– November 2005
n=255
Italian-speaking patients
Inclusion from January 2001
– November 2005
n=53
German-speaking patients
Inclusion from January 2006
– April 2009
n=165
Italian-speaking patients
Inclusion from January 2006
– November 2014
n=119
Exclusion n=39
• Written language skills n=31
• Other reasons n=8 Exclusion n=9
• Written language skills
n=1(Portuguese)
• Other reasons n=8 Exclusion n=35
• Written language skills n=32
• Other reasons n=3 Exclusion n=19
• Written language skills
n=7(Portuguese)
• Compliance n=3
• Other reasons n=9
Drop out n=80
• Premature discharge n=3
• Compliance n=77 Drop out n=9
Premature discharge n=1
Compliance n=8 Drop out n=67
• Premature discharge n=9
• Compliance n=57 Drop out n=39
• Premature discharge n=4
• Compliance n=35
Complete data included in
analysis
n=136
Complete data included in
analysis
n=35
Complete data included in
analysis
n=63
129
Complete data included in
analysis
n=61
Fig. 1. Flow chart of study participants.
capacity of 6 patients per treatment group in the ZISP. Only
one group of patients was treated at a time. Twice a year, the
programme was held in Italian for the ISP. This means that 2
groups with ISP and 10 groups with GSP were treated per year.
This resulted in an unbalanced number of patients included in
sample 1 with the same observation duration (Fig. 1). In order
to obtain almost equal numbers of patients, different observation
periods were chosen in sample 2.
Measures
Sociodemographic and potentially confounding parameters,
such as age, sex, occupation, living conditions, sports, and
formal education, were recorded at admission to the clinic on
a standardized form used previously in many studies (6). Co-
morbidities were retrieved from the medical history.
The SF-36 comprehensively measures the dimensions of
quality of life, physical, mental and psychosocial health (21).
This instrument contains 36 items in 8 health domains: bodily
pain, physical functioning, role physical, general health, vita-
lity, social functioning, role emotional, and mental health. It
is a commonly used measurement for the self-assessment of
health-related quality of life in chronic pain diseases, such as
fibromyalgia (22). It has already been used to assess the efficacy
of interventions in rheumatology, physiotherapy, drug treatment,
tai chi and many others (22). The validated German version was
used for the GSP (23). In sample 1, version 1 (21) was used and
in sample 2, version 2 (20). For the ISP, the validated Italian
version was used (24).
Analysis
Patients from sample 1 were assessed at baseline (T=0),
discharge (T1; short-term), i.e. 4 weeks after entry, 3 months
after entry (T2; mid-term), and 6 months after entry (T3; mid-
term). Patients from sample 2 were assessed at baseline (T=0),
discharge (T1; short-term), i.e. 4 weeks after entry, and 12
months after entry (T4; mid-term).
SF-36 scores were transformed into scales ranging from 0
(“maximal symptoms or limitation”) to 100 (“no symptoms or
limitation”) to ease comparison of the descriptive data (25). The
specific “missing rules” of the instrument had to be fulfilled for
determination of the scales. This means that at least 50% of the
items had to be completed for each of the SF-36 scales (25).
Sociodemographic and disease-relevant frequency data
were compared by the χ 2 test and continuous data by the non-
parametric Wilcoxon test. Changes on the SF-36 scales between
baseline and follow-up were quantified by multivariate stan-
dardized mean differences (SMD) (26). For each SF-36 score,
stepwise multivariate linear regression was used to model the
individual score changes (baseline to follow-up) as dependent
variables. The same independent variables were used for all
scales in both samples: group allocation (1=GSP, 0=ISP), ba-
seline score, and sex and education (27). The last 3 variables
are well-known as potential confounders for the score changes
between baseline and follow-up. The number of confounders is
limited by the number of patients in the smallest group/10 (28).
The coefficient/slope of the group allocation variable was then
equal to the adjusted score difference and was used to calculate
the multivariate SMDs (26).
The SMD equals the difference of the mean score changes
(baseline to follow-up) between the 2 groups (GSP and ISP)
divided by the pooled standard deviation of the score changes
(baseline to follow-up) of the 2 compared groups (26). The
pooled variance equals the mean of the 2 score change varian-
ces, which is weighted by the number of patients. Intervals for
95% confidence (95% CI) for the SMD and t-test based type I
errors (p) for testing SMD > 0.00 (zero outside of the 95% CI)
J Rehabil Med 51, 2019