Mental Health and Physical Activity 12 (2017) 73e82

Contents lists available at ScienceDirect

Mental Health and Physical Activity

journal homepage: www.elsevier.com/locate/menpa

Join The Walk?: Short-term and follow-up effects of a 10-week

walking intervention in patients with a mental disorder
Jari Vanroy a, *, Jan Seghers a, An Bogaerts b, Anne Wijtzes a, Filip Boen a
a
KU Leuven, Department of Kinesiology, Tervuursevest 101, 3001 Leuven, Belgium
b
KU Leuven, Faculty of Kinesiology and Rehabilitation Sciences, Tervuursevest 101, 3001 Leuven, Belgium

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: This study investigated the short-term and follow-up effects of a 10-week walking intervention
Received 1 July 2016 on physical tness, physical activity, anxiety and depression in patients with a mental disorder.
Received in revised form Method: A 2
3 repeated measures design was used, with condition (intervention/control) as between-
6 January 2017
subject variable. Participants in both conditions (intervention, n 91; control, n 44) were adult
Accepted 27 February 2017
Available online 6 March 2017
members of the Flemish federation for Sports and Recreation in Mental Health Care. They were diag-
nosed with a mental disorder such as mood (44.4%), psychotic (28.7%) and/or anxiety disorder (21.3%).
Physical tness, physical activity, anxiety and depression levels were assessed before (pre) and after
Keywords:
Exercise
(post) the intervention and six months later (follow-up). The intervention was based on Self-
Intervention Determination Theory and embedded in existing associations of the Flemish federation for Sports and
Pedometer Recreation in Mental Health Care. Participants in the intervention received a personalized pedometer-
Physical activity based walking schedule, a weekly group walk and guidance by a walking coach.
Psychological illness Results: There were no signicant interaction effects between time and condition for any of the out-
comes. However, from pre to post, self-reported physical activity increased signicantly across conditions
(p < 0.05).
Conclusions: The ndings suggest that patients with a mental disorder who engage in a physical activity
intervention study, increase their self-reported physical activity levels at short term.
ClinicalTrialsID: NCT02079012.
2017 Elsevier Ltd. All rights reserved.

The burden of mental disorders is growing, with signicant
health and social consequences worldwide (World Health
Organization, 2016). Patients with a mental disorder have a
shorter life-expectancy (Thornicroft, 2013) and a reduced quality of
life. This is not only a direct consequence of their condition but also
indirectly caused through stigmatization (Corrigan & Watson,
2002) and through comorbid physical illness (Newcomer, 2007).
For example, meta-analyses have shown that a severe mental dis-
order heightens the risk of metabolic syndrome (Vancampfort et al.,
2015b) and of diabetes type 2 (Vancampfort et al., 2016).
There are many types of mental disorders, such as anxiety dis-
order and depression. Worldwide, approximately one out of six
people suffers from any anxiety disorder during his lifespan
(Kessler et al., 2009) and currently 350 million people are affected
by depression (World Health Organization, 2015). Notably, other
* Corresponding author.
E-mail address: Jari.Vanroy@kuleuven.be (J. Vanroy).
types of mental disorders such as substance-abuse (National
Alliance of Mental Ilness, 2016) or psychosis (Veras, do-
Nascimento, Rodrigues, Guimara ~es, & Nardi, 2011), are often
related to feelings of anxiety and/or depression.
Conventional treatment of mental disorders by means of
medication is often accompanied by numerous side effects
(National Institute of Mental Health, 2016) and it is therefore
important to explore alternative ways of treatment. Amongst
others, physical activity (PA) has been suggested as an alternative
strategy in the treatment process (Wolff et al., 2011). PA has been
associated with a number of benecial outcomes, including
cardiorespiratory tness and lower all-cause mortality (Lee et al.,
2011), physical health (e.g., decreased risk of diabetes type 2), but
also mental health (Centers for Disease Control and Prevention,
2015). Antidepressant and anxiolytic effects of PA have been pro-
posed (Stro hle, 2009). A recent meta-analysis indicated that exer-
cise has a large and signicant antidepressant effect in people with
depression and that this effect had been underestimated due to
publication bias (Schuch et al., 2016). Research even suggests a

http://dx.doi.org/10.1016/j.mhpa.2017.02.005
1755-2966/ 2017 Elsevier Ltd. All rights reserved.
74 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82
protective effect of PA on the development of certain mental dis-
orders such as an anxiety disorder (Stro hle et al., 2007).
Nevertheless, people with a mental disorder are generally not
physically active. In a sample of 165 adults with mild to moderate
symptoms of anxiety/depression, a low compliance to PA guide-
lines was shown as well as a negative relationship between
depressive symptoms and light PA (Helgado ttir, Forsell, & Ekblom,
2015). Another study showed that patients with severe mental
illness were more likely to perform no PA at all than a matched
sample from the general population (Daumit et al., 2005). In
addition, inverse relationships between depressive symptoms and
physical tness have been demonstrated (Sui et al., 2010). None-
theless, research on PA as a treatment for patients with a mental
disorder is sparse (Zschucke, Gaudlitz, & Stro hle, 2013).
The current study focuses on one particular type of PA, i.e.
walking. This low-cost activity can be performed alone or in group
and at a wide range of skill levels and intensities, especially at lower
intensity. This is important because research has indicated a lack of
light intensity PA in people with mild to moderate depression and/
or anxiety disorder symptoms (Helgado ttir et al., 2015) as well as a
lack of any PA in people with a severe mental disorder (Daumit
et al., 2005). Walking is a form of PA that is preferred by and
appropriate for people with a mental disorder (Daumit et al., 2005;
Richardson et al., 2005; Soundy, Muhamed, Stubbs, Probst, &
Vancampfort, 2014a). A systematic review concluded that walking
can alleviate depressive symptoms in people with depression (R.
Robertson, A. Robertson, Jepson, & Maxwell, 2012). Another sys-
tematic review concluded that walking can decrease BMI at short
term in people with a schizoaffective disorder (Soundy et al.,
2014a). However, both of these reviews included only a limited
number of studies. Moreover, these studies were conducted in se-
lective samples and offered a restricted view of outcomes that are
relevant to the clinical practice.
Walking interventions seem a viable means to promote walking,
although their clinical benets are uncertain (Ogilvie et al., 2007).
Pelssers et al. (2013) showed positive effects of a 10-week walking
intervention on physical tness, PA and anxiety levels among
healthy people aged 55 years and over. This intervention, called
Every Step Counts! (ESC), was embedded in local associations
from a Flemish sociocultural organization for elderly. The inter-
vention consisted of a personalized pedometer-based walking
schedule, a weekly group walk and guidance by a walking coach.
Because of the positive effects, the wide implementation (39 as-
sociations participated) and the social organization structure of this
previous walking study (i.e., ESC), the same principles were applied
to the walking intervention in the current study, called Join The
Walk? (JTW).
JTW was in essence similar to ESC but was attuned to the current
target population (i.e., patients with a mental disorder). In fact,
both ESC and JTW were in line with the basic principles from Self-
Determination Theory (Deci & Ryan, 1985). Self-Determination
Theory (SDT) has been suggested as a guiding framework for PA
interventions in mental rehabilitation (Vancampfort & Faulkner,
2014). According to SDT, a climate that fosters the needs for au-
tonomy, relatedness and competence is conducive to autonomous
motivation, well-being, and behavioral persistence. Autonomy re-
fers to a sense of volition and authorship. This need can be satised
when people make meaningful choices. Relatedness refers to a
feeling of connection to a social network. This need can be satised
when people genuinely involve with others. Competence refers to a
perception of control over the environment. This need can be
satised when people envision attainable targets. Autonomous
motivation has been shown to relate positively to walking and to
moderate and vigorous PA in patients with a mental disorder
(Vancampfort, Stubbs, Venigalla, & Probst, 2015c). In addition, a
multicenter cross-sectional study in people with affective disorders
indicated signicant positive correlations between autonomous
regulation in exercise and PA levels, walking amounts, and positive
affect. Negative correlations were observed between autonomous
regulation and negative affect. The opposite pattern of correlations
was found between amotivation and these outcomes (Vancampfort
et al., 2015a).
The aim of the current study was to investigate the short-term
and follow-up effects of JTW. Based on the results from Pelssers
and colleagues (Pelssers et al., 2013), we hypothesized that patients
with a mental disorder who participated in a 10-week walking
intervention similar to ESC, would improve on physical tness, PA,
levels of anxiety (lower feelings of anxiety) and levels of depression
(lower feelings of depression), compared with patients in a control
condition.
1. Method
1.1. Design
This was a non-randomized controlled trial. Participants were
adult members of the Flemish federation for Sports and Recreation
in Mental Health Care (viz., with a diagnosed mental disorder). A
2
3 repeated measures design was used. Condition (intervention/
control) served as between-subject variable. Participants in the
intervention condition (IC) took part in a 10-week intervention,
whereas participants in the control condition (CC) did not. Partic-
ipants in the CC were not recommended or prohibited any (phys-
ical) activities. Physical tness, PA, anxiety and depression were
assessed pre intervention (spring 2014), post intervention (10
weeks after pre) and six months after the end of the intervention
(i.e., follow-up).
In order not to overload participants with information, the
baseline measurements (pre-test) in the IC were conducted in two
parts, separated by a one-week interval (two-week in one associ-
ation; see further). The separation was necessary because the
intervention was explained immediately after the second part,
which included the pedometer-based six-minute walk test (see
further). Hence, the second baseline measurement formally
marked the start of the intervention.
1.2. Recruitment
1.2.1. Associations
JTW was embedded in associations afliated with the Flemish
federation for Sports and Recreation in Mental Health Care
(henceforth Federation). These associations offer all kind of ergo-
therapeutic activities but they also receive support from the
Federation in particular to support the provision of their PA and
sports programs. The associations are attached to a mental health
institution, such as a daycare center or a psychiatric center, and
their members constitute patients who have been diagnosed with a
mental disorder by a physician, most of them for a longer (chronic)
period of time. The types (e.g., substance abuse, anxiety disorder,
depression, psychotic disorder), durations and degrees of severity
of the mental disorders vary. This diversity fosters a multidisci-
plinary sphere, which is currently lacking in the domain of PA in-
terventions in mental rehabilitation (Vancampfort & Faulkner,
2014).
Associations were recruited in two steps. First, the Federation
recruited associations for the IC, to ensure sufcient participation in
the project. The Federation announced the project among all as-
sociations through diverse communication channels (e.g., news-
letter, e-mail). The twelve associations that responded rst were
selected. In March 2014, a communal information session was
 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82
organized for instructors and walking coaches from these twelve
associations; the term instructor refers to a caregiver in general,
whereas the term walking coach refers to a specic function that
can be fullled by either a caregiver or a participant (see further). In
an information session, the instructors and walking coaches were
taught how to take the six-minute walk test and how to deliver the
intervention in line with basic principles from SDT and related
behavior change theories. At least one instructor or walking coach
of eleven associations attended an information session. In addition,
all associations received a manual with similar content as provided
in an information session.
One of the associations seemed not able to comply with the
protocol in several ways and was therefore excluded from this
study. In another association, the intervention started in summer
instead of spring, about two months after the baseline measure-
ments. During the week before start, the instructors from this as-
sociation organized additional moments to explain the
intervention again.
Second, the Federation recruited participants for the CC through
an e-mail among remaining associations. Participating associations
in the CC received a gadget and would be contacted for participa-
tion in a subsequent phase of the project before other associations.
Five associations enrolled in the CC. Three participants in the
control condition were only able to perform their post-test about
ve weeks after the prescheduled post-test meeting.
1.2.2. Participants
In a following phase, instructors from the enrolling associations
in both conditions invited adult members, all diagnosed with a
mental disorder, to participate in the project (up to a maximum of
10). These instructors had been asked to target their most inde-
pendent members; that is, the ones the instructors thought who
were able to deal with the information and who were capable of
walking on their own. Information about the project was also
available at the website of the Federation. A total of 135 people
started the study, 91 in the IC and 44 in the CC. The number of
participants per (division of an) association varied between four
and sixteen. All participants signed an informed consent form and
were given permission to participate by a physician. This study was
approved by the Medical Ethics Committee UZ KU Leuven
(ML10148; ClinicalTrialsID: NCT02079012).
Table 1
Intervention elements and theoretical links.
Autonomy Support Relatedness Support
Schedule & Pedometer - Choicea
- Self-monitoringb
Group Walk - Experience discussion - Attachments
- Identify PA
- Signicant others
- Social setting
Walking Coachc - Experience discussion - Information provision
- Support
a
Apart from the weekly group walk, the schedule allowed for individual decisions regarding timing (e.g., before lunch), location (e.g., xed tour) and companionship (e.g.,
alone).
b
Self-monitoring has often been used in successful walking interventions (Bird et al., 2013).
c
The walking coach resembles the Physical Activity Facilitator (Haase et al., 2010) but instead involves an insider (i.e., a more integrative approach).
75
1.3. Intervention
The general purpose of JTW was to enhance PA, physical tness
and (mental) health levels of patients with a mental disorder in a
sustained way. JTW was based on the ESC intervention (Pelssers
et al., 2013) and consisted of the following three key elements: a
pedometer-based walking schedule, a weekly group walk, and a
walking coach. Table 1 displays the links between these elements
on the one hand and SDT (Haase, Taylor, Fox, Thorp, & Lewis, 2010)
and established behavior change techniques (Abraham & Michie,
2008) on the other hand.
First, each week participants received a personal walking
schedule. This schedule was based on their baseline tness levels
(determined by a six-minute walk test e see further) and on FITT-
principles (frequency, intensity, time, type). The schedule included
recommended numbers of walks per week and steps per walk.
During the intervention, participants received a pedometer and a
booklet to record their steps.
Second, a weekly group walk was recommended per association.
Nine of the eleven participating associations in the IC were able to
organize at least eight group walks. One association did not orga-
nize any formal group walks because of other activities at the as-
sociation. Instead, participants from this association met otherwise,
whether or not in combination with a walk. In the other association
all four participants (of the total 91) dropped out before the end of
the intervention.
Third, each association appointed one or two walking coaches
on a voluntary basis. Walking coaches' duties were to organize the
group walks and to hand out the personal walking schedules. It was
recommended that each association had a walking coach who was
also a participating member. This was the case in ve associations.
In the other six associations a caregiver fullled the function of
walking coach. Nonetheless, participating member-walking
coaches were often supported by an instructor too.
1.4. Measurements
1.4.1. Procedure
The assessment of all outcome variables occurred at a location of
the associations, took approximately one hour and was conducted
three times in each condition (pre, post and follow-up). Guided by a
researcher and/or instructor, participants performed a physical
tness test (i.e., six-minute walk test) and lled out a questionnaire.
Competence Support Behavior Change Technique
- Achievable steps - Graded tasks
- Progress monitoring - Goal setting
- Self regulation - Goal review
- Performance feedback
- Self-monitoringb
- Mastery focus - Encouragement
- Modeling
- Social comparison
- Social support
- Dening PA - Barrier identication
- Success/failure reection - Encouragement
- Identication
- Information provision
- Instruction
- Modeling
76 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82
The questionnaire could be lled out individually or with assistance
if participants for example needed help with interpretation and/or
memory. Answers to objective questions (e.g., demographics) were
briey discussed between researchers and instructors if necessary.
1.4.2. Demographics
Age, sex, height and weight (BMI), daily smoking and type of
mental disorder diagnosed (semi-open) were determined by the
questionnaire. Information about medication stability (last four
weeks) and comorbid substance dependence (last six months) was
delivered by an instructor and/or a physician.
1.4.3. Physical tness
The six-minute walk test was used to assess physical tness,
expressed as distance walked in six minutes. This test has been
shown to be reliable in patients with schizophrenia (Vancampfort
et al., 2011). More specically, participants had to walk as many
lengths of 20 m as possible within six minutes (Butland, Pang,
Gross, Woodcock, & Geddes, 1982). Total distance walked was ob-
tained by multiplying the number of steps they made during the
test, as registered by a pedometer, with their step length, as
determined in advance of the six-minute walk test (Pelssers et al.,
2013). This approach was preferred as two participants could be
tested at the same time, which reduced the total time per mea-
surement. Whenever two participants performed the six-minute
walk test simultaneously, they started at opposite sides. Some
test scores were excluded from analysis because of extremely
confounding circumstances or characteristics, such as a large
physical impairment (pre: n 5; post: n 2; follow-up: n 1).
1.4.4. Physical activity
An adapted version of the Godin-Shepard Leisure-Time Exercise
Questionnaire was used to assess PA (Godin, 2011). Participants had
to report how many sessions of at least fteen minutes they had
been performing light, moderate and vigorous PA during the pre-
vious week. In contrast to the original questionnaire, the ques-
tionnaire in the current study did not explicitly ask for typical
activities (unless the previous week had been extremely uncom-
mon) or leisure activities. This approach was adopted to facilitate
recall and to prevent ambiguity about leisure time categorization.
To illustrate, patients with a mental disorder can easily classify
ergo-therapy as both work and leisure time. A composite PA score
was created by summing the three partial scores (light, moderate
and vigorous PA), multiplied by their relative weight (3, 5 and 9
respectively; Godin, 2011).
1.4.5. Anxiety level
To avoid extensive questioning, ve items were used to assess
general anxiety level. Participants had to indicate on a 5-point
Likert scale to what extent these items applied to them in general
(1 never; 3 sometimes; 5 very often). These items were based
on the trait version of the State-Trait Anxiety Inventory
(Spielberger, 1983). They were selected on the basis of discriminant
face validity and pilot tested for comprehensibility. Because of low
internal consistencies at pre-, post- and follow-up-test (Cronbach's
alpha's ranged between 0.49 and 0.70), two positively formulated
items were excluded (feeling safe and taking decisions easily). In-
ternal consistencies for the remaining scale were acceptable
(Cronbach's alpha's ranged between 0.62 and 0.74). The three items
retained in this scale were all negatively formulated (i.e., higher
scores indicate higher anxiety levels) and probably tapped more
into an anxiety component of negative affect rather than a
depression component, in contrast with the two items that had
been excluded (Bieling, Antony, & Swinsonc, 1998). An example of a
retained item is: In general I feel nervous, restless. The mean of
the remaining three items was calculated for further analyses.
1.4.6. Depression level
To avoid extensive questioning, ve items were used to assess
depression level during the last week. Participants had to indicate
on a 5-point Likert scale to what extent these items applied to them
during the last week (1 never; 3 sometimes; 5 very often).
The items were based on communal themes in the Beck Depression
Inventory (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), the
Hamilton Depression Rating Scale (Hamilton, 1960) and the 16-
item self-report Quick Inventory of Depressive Symptomatology
(Rush et al., 2006). They were selected on the basis of discriminant
face validity and pilot tested for comprehensibility. Internal con-
sistencies at pre-, post- and follow-up-test were acceptable
(Cronbach's alpha's ranged between 0.69 and 0.73). An example
item is: Last week I felt sad. The mean of the items was used for
further analyses.
1.4.7. Covariates
Several variables were assessed as covariates. First, relevant
demographic characteristics from the baseline measurement were
taken into account (age, sex, institutionalization, BMI, daily smok-
ing, medication stability, comorbid substance dependence and
anxiety/mood disorder). Second, because research has shown that
climatic circumstances play a role in people's PA behavior (Chan &
Ryan, 2009), three categories of climatic circumstances during the
week before PA-assessment were determined: (a) day length in
hours daylight (in Brussels), (b) mean day temperature in degrees
Celsius and (c) day precipitation in liters per square meter (both in a
region nearby PA-assessment). Day length was based on data from
the Royal Observatory of Belgium while day temperature and pre-
cipitation were based on data from the Royal Meteorological
Institute of Belgium. Weekly means were taken. Third, two cate-
gories of medication intake were determined: antipsychotic
medication and antidepressant medication. This information was
based on reports from participants, instructors and physicians and
summarized into a dichotomous measure (yes/no). Other medica-
tion with a mood stabilizing effect (e.g., lithium carbonate) or
calming effect (e.g., Valium), were not included in the summary
measure.
1.5. Analyses
SPSS 23 was used for statistical analyses. The signicance cri-
terion (p < 0.05), was set two-sided (without correction for mul-
tiple testing). All outcome variables (i.e., physical tness, PA,
anxiety and depression) were considered continuous. Progress in
these variables was predicted over two time periods: (1) from pre
to post and (2) from pre to follow-up. Because both time periods
were based on two time points, progress was assumed linear and
assessed by means of linear mixed models. This differentiated
approach was adopted because of the differences in time intervals
at one hand and specic theoretical predictions on the other (i.e.,
SDT aims at sustainable behavior change). Condition (IC/CC), time
(pre/post and pre/follow-up) and the covariates were treated as
(repeated) xed factors. Covariates were selected outcome-
specically (e.g., anxiety disorder as covariate for anxiety level).
Unstructured covariance matrices and restricted maximum likeli-
hood estimations were used. Signicant interaction effects (con-
dition x time) indicate an intervention effect.
Two extremely high PA scores at post-test and one extremely
high score at follow-up-test were excluded for further analysis.
These outliers were all higher than 152, which is the equivalent of
six out of seven days with a bout of light, moderate and vigorous PA
and which stands in sharp contrast with the median scores (30).
 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82 77

Such high scores are probably caused by misinterpretation of the Table 2

question (i.e., distribution of one longer PA session into separate Demographic baseline characteristics.

parts). IC (n 91) CC (n 44) p

Calculations of skewness divided by standard error of skewness Age (M SD years) 46.4 10.8 44.6 15.7 0.78
generally indicated that scores on physical tness and PA were Sex (% males) 53.8 63.4 0.35
right-skewed. Consequently, these scores were replaced by their Institutionalization (% opena) 89.8 40.5 0.00*
square root (Tukey, 1977). These transformations yielded satisfac- BMI (M SD kg/m2) 29.8 6.8 27.2 4.6 0.02*
Daily smoking (M SD n cigarettes) 11.3 12.4 12.0 11.6 0.65
tory improvements in symmetry such that all but one (out of six)
Medication stability (%) 95.5 58.3 0.00*
corresponding quotients resulted in z-values below the threshold Comorbid substance dependence (%) 12.5 45.9 0.00*
of 1.96 (Kim, 2013). Only for physical tness at post-test, the cor- Anxiety disorder (%) 26.7 9.8 0.04*
responding quotient remained above this threshold, but to an Mood disorder (%) 47.1 39.0 0.48
Psychotic disorder (%) 27.6 31.0 0.68
acceptable level of 2.21 (instead of 3.63).
*Signicantly different between IC and CC (p < .05).
Note 1: percentages refer to valid cases. The maximum number of missing cases was
2. Results
25 in the IC and 8 in the CC (e.g., due to unclear medical reports).
Note 2: other demographic variables such as work status were less relevant in this
2.1. Sample sample because of participants' social situation.
a
Open institutionalization means that patients are free to choose what times they
Fig. 1 shows an overview of participation rates at the three are at the institution (mostly a daycare center) and is opposed to psychiatric
connement.
measurement points, along with reasons for non-participation. The
large majority of participants took part in the pre-tests, about two
in three in the post-tests and about half in the follow-up-tests. In Demographic baseline characteristics are shown in Table 2. The
both conditions, the most frequently given reason for non- IC and CC differed signicantly at baseline on BMI (Mann-Whitney
participation was that a patient was no longer institutionalized U: p 0.02), institutionalization, medication stability, comorbid
(and hence no longer involved in the association). Lack of motiva- substance dependence (2-sided Fisher's Exact: p 0.00) and anx-
tion and test difculties were however more common in the IC than iety disorder (2-sided Fisher's Exact: p 0.04). This means that
in the CC.

Intervention condition Control condition

11 clubs, n = 91 5 clubs, n = 44

n = 85 n = 42
Pre
Non-participation reasons: (week 0) Non-participation reasons:
- lack of motivation (n = 1) - lack of motivation (n = 1)
- forbiddena (n = 3) - practical issueb (n = 1)
- practical issueb (n = 2)

n = 57 n = 32
Post
Non-participation reasons: Non-participation reasons:
(week 10)
- test was too difficult (n = 7) - test was too difficult (n = 1)
- lack of motivation (n = 4) - medical problem (n = 1)
- medical problem (n = 3) - no longer
- no longer institutionalized (n = 8)
institutionalized (n = 12) - practical issue (n = 4)
- forbidden (n = 3)
- practical issue (n = 5)

n = 53 n = 21
Follow-up
Non-participation reasons: Non-participation reasons:
(week 36)
- test was too difficult (n = 6) - test was too difficult (n = 1)
- lack of motivation (n = 9) - medical problem (n = 4)
- medical problem (n = 4) - no longer
- no longer institutionalized (n = 14)
institutionalized (n = 16) - practical issue (n = 4)
- practical issue (n = 3)

Fig. 1. Test participation at the three measurement points for both conditions, with reasons for non-participation.
a. Some patients wanted to take part in the tests to skip therapy.
b. For example, on a holiday.
78 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82

patients in the IC generally had a higher BMI than patients in the
CC. Patients in the IC were also less likely to be in psychiatric
connement as well as to be dependent on comorbid substances
(e.g., alcohol). They were more likely to have a stable medication
intake as well as to be diagnosed with an anxiety disorder. With
regard to the outcome variables, patients in the IC had signicantly
lower tness levels at baseline than patients in the CC (Mann-
Whitney U: p 0.00).
2.2. Short-term effects
From pre-to post-test, none of the hypothesized interaction ef-
fects (condition x time) were signicant for any of the measured
variables (see Table 3). Despite the lack of signicant interaction
effects, a signicant main effect of condition was found for physical
tness (p 0.03): patients in the CC scored better on the six-minute
walk test than patients in the IC (on average, 70 m at pre-test and
64 m at post-test, based on the squared square root means).
Across conditions, one signicant main effect of time was found
for PA (p 0.01): PA scores increased signicantly from pre to post
(10 points on average, based on the squared square root means). An
increase of 10 points is roughly equivalent to either one additional
weekly bout of vigorous PA, two bouts of moderate PA or three
bouts of light PA. Post hoc paired samples t-tests indicated that PA
increased signicantly in both conditions (ps < 0.05).
2.3. Follow-up effects
From pre to follow-up-test, none of the hypothesized interaction
effects (condition x time) were signicant for any of the measured
variables (see Table 4). In line with the short-term effects, a sig-
nicant main effect of condition was found for physical tness
(p 0.02): patients in the CC scored better than patients in the IC
(on average, 48 m at pre-test and 60 m at follow-up-test, based on
the squared square root means).
2.4. Health-related PA
As mentioned above, no interaction effects (condition x time)
were found in the prediction of PA as a continuous variable.
Nevertheless, a PA score that only accounts for moderate and
vigorous PA (with identical weights, i.e. 5 and 9 respectively) might
be more relevant for health benets (Godin, 2011). To this end, a
cut-off value of 24 has been proposed as the health norm (Godin,
2011). For this reason, we used this cut-off score for post hoc
categorizations.
The results revealed that from pre-to post-test, only thirteen
participants in the IC (23%) and nine participants in the CC (32%)
categorically changed from not meeting the health norm to
meeting the health norm. By contrast, eight (14%) and two partic-
ipants (7%) respectively changed in the other direction. From pre-
test to follow-up-test, eight participants in the IC (22%) and three
participants in the CC (17%) changed from not meeting the health
norm to meeting the health norm. By contrast, ve (14%) and two
participants (11%) respectively changed in the other direction.
2.5. Regression to the mean
Spearman correlations revealed consistently sizeable
(rs < 0.36) and signicant (ps < 0.01) negative relationships be-
tween the baseline value of an outcome variable (e.g., PA level at
pre-test) and its corresponding difference score at both time pe-
riods (e.g., PA level at follow-up-test e PA level at pre-test).
Therefore, we performed separate regression analyses with each
of the eight difference scores as an outcome variable (e.g., PA level
at post-test e PA level at pre-test). In these analyses, the baseline
value of the outcome variable was entered as a predictor in the
model (step 1), followed by addition of the condition (IC/CC; step
2). The explained variance did not improve signicantly for any of
the difference scores after the rst step (ps > 0.05). The highest
(non-signicant) improvement in explained variance due to con-
dition was observed for the change in the physical tness level from
pre to follow-up (R2 change 3.0%).
3. Discussion
This study evaluated the short-term and follow-up effects of a
SDT-based walking intervention on physical tness, PA, anxiety and
depression among patients with a mental disorder. No signicant
interaction effects between condition and time emerged immedi-
ately after the intervention or at follow-up six months later. In
other words, contrary to our expectations the intervention did not
lead to higher levels of PA or physical tness, or to lower levels of
anxiety or depression.
Several explanations are possible for the absence of the hy-
pothesized interaction effects. First, although the current inter-
vention JTW was based on ESC, in which a similar walking
intervention was successfully implemented among people aged 55

Table 3
Short-term effects.

Condition Pre Post F condition F F

(M SD)a (M SD)a time condition
x time

Physical tness (m)b IC 504 522 4.87* 0.79 0.51

CC 559.5 558
PA IC 28 31 1.07 7.10* 0.14
(score)c CC 21.5 33.5
Anxiety level IC 2.88 1.14 2.84 1.09 0.02 0.47 0.01
(1e5)d CC 3.11 1.10 3.02 1.02
Depression level (1e5)e IC 2.70 0.96 2.60 1.00 0.22 0.49 0.27
CC 2.73 0.87 2.66 0.87

Notes: sample sizes for the four outcome variables at both measurement points ranged between fty-four and twenty-ve participants per condition; similar analyses without
covariates did not yield different signicant results.
*p < 0.05.
a
For physical tness and PA, the median of the untransformed data is given.
b
Covariates: age, sex, institutionalization, BMI (pre), daily smoking (pre), medication stability (pre) and comorbid substance dependence (pre).
c
Covariates: age, sex, institutionalization, BMI (pre), day length (post), temperature (post) and precipitation (post).
d
Covariates: age, sex, institutionalization, anxiety disorder (pre) and antipsychotic medication (pre).
e
Covariates: age, sex, institutionalization, mood disorder (pre) and antidepressant medication (pre).
 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82
Table 4
Follow-up effects.
Condition Pre
(M SD)a
Physical tness (m)b IC 505.5
CC 546.5
PA IC 28
(score)c CC 19
Anxiety level IC 3.15 1.15
(1e5)d CC 3.39 1.02
Depression level (1e5)e IC 2.65 0.96
CC 2.72 0.89
Note: sample sizes for the four outcome variables at both measurement points ranged between forty and seventeen participants per condition.
*p < 0.05.
a
For physical tness and PA, the median of the untransformed data is given.
b
Covariates: age, sex, institutionalization, BMI (pre), daily smoking (pre), medication stability (pre) and comorbid substance dependence (pre).
c
Covariates: age, sex, institutionalization, BMI (pre), day length (follow-up), temperature (follow-up) and precipitation (follow-up).
d
Covariates: age, sex, institutionalization, anxiety disorder (pre) and antipsychotic medication (pre).
e
Covariates: age, sex, institutionalization, mood disorder (pre) and antidepressant medication (pre).
f
This value was marginally signicant in this analysis (p .06) and signicant in a similar analysis without covariates (p .04).
years and over (Pelssers et al., 2013), both studies also differed in
various ways apart from the target population. ESC had a larger
participation rate (N 580) and more homogeneous walking
groups (e.g., age range). Moreover, participants in ESC went to the
associations primarily for voluntary sociocultural interaction
whereas participants in JTW went to the associations primarily as
part of their clinical treatment.
Second, the lack of intervention effects might also be explained
by the fact that the intervention did not produce PA of sufcient
intensity to have an impact on any of the outcome variables. With
regard to physical tness and PA, it should be noted that the six-
minute walk test could not capture endurance capacity and that
the PA score was largely inuenced by the relative weight of
(perceived) intensity. Moreover, pronounced improvements in the
intervention condition with respect to PA might have been
obscured by the measurement tool. In fact, the walking schedule
prescribed longer walks on fewer days as tness levels improved.
Because the PA score did not account for total activity duration but
only for the number of sessions of at least fteen minutes, such
improvements could not be captured. Furthermore, improved
tness (and enjoyment) levels could have resulted in lower
perceived intensity levels during walking, with lower self-reported
PA scores as a consequence (e.g., activities classied as vigorous
before the intervention might have been classied as moderate
after the intervention).
With regard to the mental outcome variables, evidence from
other studies is mixed. One study in Flemish adults showed that
sports participation was the only type of PA associated with both
less stress and less distress (Asztalos et al., 2009) and another study
in Belgian adults showed that, among men, vigorous intensity PA in
particular was associated with better mental health (Asztalos, De
Bourdeaudhuij, & Cardon, 2010). However, both studies were
correlational, conducted in a non-clinical sample and lacked
theoretical coherence. Alternatively, a review by Zschucke et al.
(2013) suggested possible therapeutic effects of not only higher
but also lower intensity PA in patients with a mental disorder.
Moreover, a survey among Taiwanese elderly even found a negative
relationship between exercise intensity and well-being (Lee &
Hung, 2011).
Despite the absence of the hypothesized signicant interaction
effects, it should be noted that PA levels increased signicantly
from pre-to post-test across conditions. Several explanations for
this short-term increase are possible. First, PA improvements in the
control condition are often observed in PA intervention studies
79
Follow-up F condition F time F
(M SD)a condition
x time
516 5.58* 1.40 0.94
539.5
26 3.26 0.00 0.00
24
3.16 0.89 0.20 1.14 1.15
2.94 0.88
2.55 1.00 0.00 3.70f 1.40
2.38 0.81
(Waters, Reeves, Fjeldsoe, & Eakin, 2012). In the current study, mere
enrollment and PA measurement might have generated a short-
lived interest among participants and caregivers for physical ac-
tivities. Second, the questionnaire at baseline could have raised
awareness for subsequent recall of physical activities at short term.
Third, considering that the current study took place in a clinical
population, short-term improvements in health-related behaviors
such as PA may occur due to medical care and regression to the
mean.
In addition to the short-term increase in PA levels, depression
levels also decreased (nearly) signicantly from pre-to follow-up-
test across conditions (on average 0.18 on a 1e5 scale; p 0.06
and without covariates, p 0.04). Decreases in depressive symp-
toms in the non-intervention condition are often observed in PA
intervention studies among patients with depression (Stubbs et al.,
2016). The decreases in the current study are probably not mainly
caused by an antidepressant effect of PA; in fact, the time lag be-
tween short term (time effect on PA) and follow-up (time effect on
depression) is rather long and post hoc analyses indeed revealed a
very small and non-signicant correlation between short-term PA
difference scores and follow-up depression level difference scores
(r 0.05; p > 0.05). Alternative explanations such as medical care,
regression to the mean and enhanced social interactions due to
study participation seem more likely.
Furthermore, anxiety levels did not show a similar decrease,
even though the absolute mean values for anxiety levels were
similar to depression levels. It should be noted that the discrepancy
in mean pre-levels of participants who completed pre- and post-
tests compared with participants who completed pre- and
follow-up-tests, was negligible for depression levels (<0.1), but not
for anxiety levels (about 0.3). Moreover, anxiety levels were
construed and measured in a more general way, referring to more
stable tendencies. This suggests that patients with lower pre-levels
of anxiety might have dropped out, particularly because they were
no longer institutionalized. Post hoc descriptive analyses revealed
that almost one in three (9/31) patients with an anxiety pre-level of
2 or less dropped out because of non-institutionalization.
3.1. Strengths and limitations
This study had several strengths. First, the intervention took
place in a multidisciplinary sphere of real-life settings that offered
PA for patients with a mental disorder, which enhanced the
ecological validity. Second, practical and theoretical insights were
80 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82
combined with empirical evidence from a previous walking inter-
vention among people aged 55 years and over and tailored to the
target group of patients with a mental disorder. Third, the study
included a follow-up. Finally, multiple covariates were accounted
for in the analyses.
A number of limitations should be taken into account for a
correct interpretation of the results. First, because of the lack of
randomization, the relatively modest sample size and the nature of
the linear mixed model, the results might have been confounded by
regression to the mean. The ndings from the regression analysis
indeed supported this statement as well as the conclusion that the
intervention did not produce the expected effects.
Second, PA was assessed through self-report. This approach was
adopted primarily because of practical reasons (including low cost
and easy administration). However, the usefulness of self-report PA
measures in patients with a mental disorder has been questioned
because of recall bias, lack of validation, comparability with other
measures, and a narrow focus on structured activities (Soundy,
Roskell, Stubbs, & Vancampfort, 2014b). In addition, PA estima-
tions might have been affected by participants' characteristics such
as the type of mental disorder (e.g., mood-related) or the physical
tness level. For example, a study showed that tter individuals
tended to over-report PA more than lesser t ones (Tomaz, Lambert,
Karpul, & Kolbe-Alexander, 2014). Furthermore, in the current
study participants were assisted by an instructor to facilitate recall
and interpretation. The presence of an instructor could have caused
over-reporting due to social desirability bias (from participants as
well as from instructors themselves). Finally, engagement in a PA
intervention study might have increased awareness of subsequent
physical activities. Hence, there is a need for objective PA measures
(e.g., pedometer, accelerometer) in the current population.
However, objective measures can lead to issues of compliance
(e.g., consistent wearing) and complexity (e.g., correct wearing).
Moreover, in patients with a mental disorder the Godin-Shepard
Leisure-Time Exercise Questionnaire has been shown to have
acceptable test-retest reliability as well as concurrent validity with
regard to vigorous and total PA (Soundy et al., 2014b). Furthermore,
the inuence of social desirability bias in PA increases in the current
study should not be overestimated, given that the time intervals
between two consequent measurements were rather large to
memorize ones responses to the previous questionnaire. Moreover,
participants as well as instructors were not aware of the way in
which the PA scores were calculated.
A third limitation is that the physical tness scores (i.e., distance
walked in six minutes) were calculated by multiplication of the
number of steps with estimated step length. This method biases
results at the expense of those people who can retain their step
length most, probably the more procient ones. However, the
alternative method that determines distance by the count of
lengths is difcult in practice and favors people who know best
how to walk straight and turn in a sharp manner.
Fourth, the lack of randomization might have blurred inter-
vention effects. In particular, patients in the CC scored consistently
higher for physical tness than patients in the IC. Patients in the CC
were also younger, had lower BMI, more of them were male and
more of them were in psychiatric connement (as a proxy for
structured healthy lifestyles), compared with patients in the IC. The
CC also included fewer patients with a mood or anxiety disorder,
but more patients with comorbid substance dependence. It should
be noted though that these factors were included as covariates.
However, other relevant covariates were either not measured (e.g.,
food, sleep, full medication prole) or not included in the model
because of complexity and/or lack of cases (e.g., association as a
level with intraclass correlations).
Fifth, the drop-out rate was fairly high and reduced the sample
size. Moreover, the reasons for non-participation suggest that drop-
out was not completely random. Both positive (i.e., non-
institutionalization) and negative (i.e., test difculties, lack of
motivation and medical problems) reasons were seen in both
conditions. The combination of a small sample size and a large
heterogeneity lowers the statistical power. It should be noted
however that a review on PA interventions among people with
mental illness reported that the sample size of the current study
was comparable with sample sizes from similar studies
(Rosenbaum, Tiedemann, Sherrington, Curtis, & Ward, 2014). In
fact, only 4 out of 38 studies in the review had a sample size larger
than 100 (i.e., Ns 361, 202, 136 and 128). The other 34 studies had
a sample size between 88 and 10, with a median value of N 39
and a mean value of N 42.94 (SD 22.49).
Sixth, the ndings from the current study cannot be generalized
to the population of people with a mental disorder because of the
sample specicity. More specically, the sample consisted of pa-
tients who were part of a specic group and most of whose prob-
ably were already interested in PA. Nonetheless, the sample was
realistic in the sense that it matched the setting and the
intervention.
Finally, the sample included divergent proles. Although this
variability in proles adds to the ecological validity and the
multidisciplinary character of a project, it might also impact on the
study outcomes and complicate revealing functional mechanisms.
3.2. Conclusions and future research
The current study did not demonstrate any effects of a walking
intervention in patients with a mental disorder on physical tness,
PA, levels of anxiety and levels of depression. However, participants
across both conditions showed increased self-reported PA levels
immediately after the intervention. This suggests that merely
measuring PA levels within the context of a project can be a suf-
cient trigger for patients with a mental disorder to enhance their PA
at short term.
Future research should clarify the optimal PA types and in-
tensities in mental rehabilitation. A rst step towards this clari-
cation can be to diversify the PA interventions by type of mental
disorder and by type of institutionalization. For example, a PA
intervention quadrant can be used with anxiety/mood disorder on
the one axis and open/closed institutionalization on the other. A
second step can be to delineate the improvement margins delimi-
ted by the medication prole with concurrent side effects. In this
pursuit, physiological and psychological pathways should be stated
and distinguished a priori. For example, it would be interesting to
know how PA, whether or not in concurrence with medication, can
contribute to relaxation on the one hand, which would probably
benet patients with anxiety, and to energization on the other
hand, which would probably benet patients with depression.
Finally, although these proposed steps for further renement/
analysis are necessary to improve our understanding about the role
of PA interventions in mental rehabilitation, the broader picture
should not be ignored. This requires an investigation of qualitative
information (e.g., participant satisfaction) and of the possibility to
implement the intervention on a larger scale (e.g., cost-
effectiveness). Furthermore, more in-depth information about the
different approaches between associations and walking coaches
could enhance the knowledge about critical success factors.
4. Declaration of conicting interests
The authors declare no conicting interests.
 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82
Role of the funding source
This study was supported by the Flemish Federation for Sport
and Recreation in Mental Health Care, which assisted in study
design and data collection.
Acknowledgements
We thank all participants and caregivers for their willingness,
patience and support. Furthermore we are grateful for the oppor-
tunities and aid from the Flemish Federation for Sport and Recre-
ation in Mental Health Care, the Royal Meteorological Institute of
Belgium and the Royal Observatory of Belgium.
References
Abraham, C., & Michie, S. (2008). A taxonomy of behavior change techniques used in
interventions. Health Psychology, 27(3), 379e387.
Asztalos, M., De Bourdeaudhuij, I., & Cardon, G. (2010). The relationship between
physical activity and mental health varies across activity intensity levels and
dimensions of mental health among women and men. Public Health Nutrition,
13(8), 1207e1214.
Asztalos, M., Wijndaele, K., De Bourdeaudhuij, I., Philippaerts, R., Matton, L.,
Duvigneaud, N., et al. (2009). Specic associations between types of physical
activity and components of mental health. Journal of Science and Medicine in
Sport, 12(4), 468e474.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory
for measuring depression. Archives of General Psychiatry, 4, 561e571.
Bieling, P. J., Antony, M. M., & Swinsonc, R. P. (1998). The state-trait anxiety in-
ventory, trait version: Structure and content re-examined. Behaviour Research
and Therapy, 36, 777e788.
Bird, E. L., Baker, G., Mutrie, N., Ogilvie, D., Sahlqvist, S., & Powell, J. (2013). Behavior
change techniques used to promote walking and cycling: A systematic review.
Health Psychology, 32(8), 829e838.
Butland, R. J. A., Pang, J., Gross, E. R., Woodcock, A. A., & Geddes, D. M. (1982). Two-,
six-, and 12-minute walking tests in respiratory disease. British Medical Journal,
284, 1607e1608.
Centers for Disease Control and Prevention. (2015). The benets of physical activity.
http://www.cdc.gov/physicalactivity/basics/pa-health/ Accessed 01.07.16.
Chan, C. B., & Ryan, D. A. (2009). Assessing the effects of weather conditions on
physical activity participation using objective measures. International Journal of
Environmental Research and Public Health, 6(10), 2639e2654.
Corrigan, P. W., & Watson, A. C. (2002). Understanding the impact of stigma on
people with mental illness. World Psychiatry, 1(1), 16e20.
Daumit, L. G., Goldberg, W. R., Anthony, H. C., Dickerson, B. F., Brown, B. C.,
Kreyenbuhl, B. J., Dixon, B. L. (2005). Physical activity patterns in adults with
severe mental illness. Journal of Nervous and Mental Disease, 193, 641e646.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behaviour. New York: Plenum.
Godin, G. (2011). The Godin-shephard leisure-time physical activity questionnaire.
ACSM's Health & Fitness Journal of Canada, 4(1), 18e22.
Haase, A. M., Taylor, A. H., Fox, K. R., Thorp, H., & Lewis, G. (2010). Rationale and
development of the physical activity counselling intervention for a pragmatic
TRial of Exercise and Depression in the UK (TREAD-UK). Mental Health &
Physical Activity, 3(2), 85e91.
Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosur-
gery & Psychiatry, 23, 56e62.
Helgado (2015). Physical activity patterns of people
 ttir, B., Forsell, Y., & Ekblom, O.
affected by depressive and anxiety disorders as measured by accelerometers: A
cross-sectional study. PLoS One, 10(1). http://dx.doi.org/10.1371/
journal.pone.0115894.
Kessler, R. C., Aguilar-Gaxiola, S., Alonso, J., Chatterji, J. S., Lee, S.,
Ormel, J., Wang, P. S. (2009). The global burden of mental disorders: An up-
date from the WHO World Mental Health (WMH) surveys. Epidemiologia e
psichiatria sociale, 18(1), 23e33.
Kim, H.-Y. (2013). Statistical notes for clinical researchers: Assessing normal dis-
tribution (2) using skewness and kurtosis. Restorative Dentistry & Endodontics,
38(1), 52e55.
Lee, Y.-J., & Hung, W.-L. (2011). The relationship between exercise participation and
well-being of the retired elderly. Aging & Mental Health, 15(7), 873e881.
Lee, D.-C., Sui, X., Ortega, F. B., Kim, Y.-S., Church, T. S., Winett, R. A., Blair, S. N.
(2011). Comparisons of leisure-time physical activity and cardiorespiratory
tness as predictors of all-cause mortality in men and women. British Journal of
Sports Medicine, 45, 504e510.
National Alliance of Mental Ilness. (2016). Dual diagnosis. https://www.nami.org/
Learn-More/Mental-Health-Conditions/Related-Conditions/Dual-Diagnosis/
Accessed 01.07.16.
National Institute of Mental Health. (2016). Mental health medications. http://www.
nimh.nih.gov/health/topics/mental-health-medications/index.shtml/ Accessed
01.07.16.
81
Newcomer, J. W. (2007). Metabolic syndrome and mental illness. The American
Journal of Managed Care, 13(Suppl. 7), 170e177.
Ogilvie, D., Foster, C. E., Rothnie, H., Cavill, N., Hamilton, V., Fitzsimons, C. F., et al.
(2007). Interventions to promote walking: Systematic review. British Medical
Journal, 334. http://dx.doi.org/10.1136/bmj.39198.722720.BE.
Pelssers, J., Delecluse, C., Opdenacker, J., Kennis, E., Roie, E. V., & Boen, F. (2013).
Every step counts!: Effects of a structured walking intervention in a
community-based senior organization. Journal of Aging and Physical Activity, 21,
167e185.
Richardson, C., Faulkner, G., Mcdevitt, J., Skrinar, G., Hutchinson, D., & Piette, J.
(2005). Integrating physical activity into mental health services for persons
with serious mental illness. Psychiatric Services, 56(3), 324e331.
Robertson, R., Robertson, A., Jepson, R., & Maxwell, M. (2012). Walking for
depression: A systematic review and meta-analysis. Mental Health and Physical
Activity, 5(1), 66e75.
Rosenbaum, S., Tiedemann, A., Sherrington, C., Curtis, J., & Ward, P. B. (2014).
Physical activity interventions for people with mental illness: A systematic
review and meta-analysis. The Journal of Clinical Psychiatry, 75(9), 964e974.
Rush, A. J., Trivedi, M. H., Ibrahim, H. M., Carmody, T. J., Arnow, B.,
Klein, D. N., Keller, M. B. (2006). The 16-item Quick Inventory of Depressive
Symptomatology (QIDS) Clinician Rating (QIDS-C) and Self-Report (QIDS-SR): A
psychometric evaluation in patients with chronic major depression. Biological
Psychiatry, 54, 573e583.
Schuch, F. B., Vancampfort, D., Richards, J., Rosenbaum, S., Ward, P. B., & Stubbs, B.
(2016). Exercise as a treatment for depression: A meta-analysis adjusting for
publication bias. Journal of Psychiatric Research, 77, 42e51.
Soundy, A., Muhamed, A., Stubbs, B., Probst, M., & Vancampfort, D. (2014a). The
benets of walking for individuals with schizophrenia spectrum disorders: A
systematic review. International Journal of Therapy and Rehabilitation, 21(9),
410e420.
Soundy, A., Roskell, C., Stubbs, B., & Vancampfort, D. (2014b). Selection, use and
psychometric properties of physical activity measures to assess individuals with
severe mental illness: A narrative synthesis. Archives of Psychiatric Nursing,
28(2), 135e151.
Spielberger, C. D. (1983). Manual for the State-trait anxiety Inventory STAI (Form Y).
Palo Alto, CA: Mind Garden.
hle, A. (2009). Physical activity, exercise, depression and anxiety disorders.
Stro
Journal of Neural Transmission, 116, 777e784.
hle, A., Ho
Stro er, M., Hildegard, P., Mller, A.-G., Hoyer, J., Wittchen, H.-U., et al.
(2007). Physical activity and prevalence and incidence of mental disorders in
adolescents and young adults. Psychological Medicine, 37, 1657e1666.
Stubbs, B., Vancampfort, D., Rosenbaum, S., Ward, P. B., Richards, J., Ussher, M., et al.
(2016). Challenges establishing the efcacy of exercise as an antidepressant
treatment: A systematic review and meta-analysis of control group responses in
exercise randomised controlled trials. Sports Medicine, 46(5), 699e713.
Sui, X., Laditka, J. N., Church, T. S., Hardin, J. W., Chase, N., Davis, K., et al. (2010).
Prospective study of cardiorespiratory tness and depressive symptoms in
women and men. Journal of Psychiatric Research, 43(5), 546e552.
Thornicroft, G. (2013). Premature death among people with mental illness. British
Medical Journal, 346(7909), 7.
Tomaz, S. A., Lambert, E. V., Karpul, D., & Kolbe-Alexander, T. L. (2014). Cardiovas-
cular tness is associated with bias between self-reported and objectively
measured physical activity. European Journal of Sport Science, 16(1), 149e157.
Tukey, J. W. (1977). Exploratory data analysis. Reading, MA: Addison-Wesley.
Vancampfort, D., Correll, C. U., Galling, B., Probst, M., De Hert, M.,
Ward, P. B., Stubbs, B. (2016). Diabetes mellitus in people with schizophrenia,
bipolar disorder and major depressive disorder: A systematic review and large
scale meta-analysis. World Psychiatry, 15(2), 166e174.
Vancampfort, D., & Faulkner, G. (2014). Physical activity and serious mental illness:
A multidisciplinary call to action. Mental Health and Physical Activity, 7(3),
153e154.
Vancampfort, D., Madou, T., Moens, H., De Backer, T., Vanhalst, P.,
Helon, C., Probst, M. (2015a). Could autonomous motivation hold the key to
successfully implementing lifestyle changes in affective disorders? A multi-
centre cross sectional study. Psychiatry Research, 228(1), 100e106.
Vancampfort, D., Probst, M., Maurissen, K., Sweers, K., Knapen, J., & De Herdt, M.
(2011). Reliability, minimal detectable changes, practice effects and correlates of
the six-minute walk test in patients with schizophrenia. Psychiatry Research,
187, 62e67.
Vancampfort, D., Stubbs, B., Mitchell, A. J., De Hert, M., Wampers, M.,
Ward, P. B., Correll, C. U. (2015b). Risk of metabolic syndrome and its com-
ponents in people with schizophrenia and related psychotic disorders, bipolar
disorder and major depressive disorder: A systematic review and meta-analysis.
World Psychiatry, 14(3), 339e347.
Vancampfort, D., Stubbs, B., Venigalla, S. K., & Probst, M. (2015c). Adopting and
maintaining physical activity behaviours in people with severe mental illness:
The importance of autonomous motivation. Preventive Medicine, 81, 216e220.
Veras, A. B., do-Nascimento, J. S., Rodrigues, R. L., Guimara ~es, A. C. A., & Nardi, A. E.
(2011). Psychotic symptoms in social anxiety disorder patients: Report of three
cases. International Archives of Medicine. http://dx.doi.org/10.1186/1755-7682-4-
12.
Waters, L., Reeves, M., Fjeldsoe, B., & Eakin, E. (2012). Control group improvements
in physical activity intervention trials and possible explanatory factors: A sys-
tematic review. Journal of Physical Activity and Health, 9, 884e895.
hle, A. (2011).
Wolff, E., Gaudlitz, K., von Lindenberger, B.-L., Plag, J., Heinz, A., & Stro
82 J. Vanroy et al. / Mental Health and Physical Activity 12 (2017) 73e82

Exercise and physical activity in mental disorders. European Archives of Psy- mediacentre/factsheets/fs396/en/. Accesed 01.07.16.
chiatry and Clinical Neuroscience, 261(Suppl 2), 186e191. hle, A. (2013). Exercise and physical activity in
Zschucke, E., Gaudlitz, K., & Stro
World Health Organization. (2015). Depression. http://www.who.int/mediacentre/ mental disorders: Clinical and experimental evidence. Journal of Preventive
factsheets/fs369/en/. Accesed 01.07.16. Medicine & Public Health, 46(Suppl. 1), 12e21.
World Health Organization. (2016). Mental disorders. http://www.who.int/