1-s2.0-S0165032706003442-main

Journal of Affective Disorders 99 (2007) 63 – 71
www.elsevier.com/locate/jad

Research report

A pragmatic, unblinded randomised controlled trial comparing an
occupational therapy-led lifestyle approach and routine GP
care for panic disorder treatment in primary care☆
Rodney A. Lambert a,⁎,1 , Ian Harvey b , Fiona Poland c
a

School of Allied Health Professions, University of East Anglia, Norwich, Norfolk, England, NR4 7TJ, UK
b
School of Medicine, Health Policy and Practice, University of East Anglia, Norwich, UK
c
School of Allied Health Professions, University of East Anglia, Norwich, UK
Received 19 June 2006; received in revised form 8 August 2006; accepted 11 August 2006
Available online 2 October 2006

Abstract
Background: Treated anxiety increased in the UK by over 30% since 1994. Medication and psychological treatment is most
common, but outcomes are sometimes poor, with high relapse rates. Lifestyle has a potential role in treatment, but is not considered
in clinical guidelines. Panic disorder is potentially influenced by lifestyle factors.
Methods: 16 week unblinded pragmatic randomised controlled trial in 15 East of England primary care practices (2 Primary Care
Trusts). Participants met DSM-IV criteria for panic disorder with/without agoraphobia. Follow-up at 20 weeks and 10 months.
Control arm, unrestricted routine GP care. Trial Arm, Occupational therapy-led lifestyle treatment comprising: lifestyle review of
fluid intake, diet pattern, exercise, caffeine, alcohol and nicotine; negotiation of positive lifestyle changes; monitoring and review
of impact of changes. Primary outcome measure: Beck Anxiety Inventory.
Data analysis: Intention-to-treat analysis provided between-group comparisons using analysis of co-variance. Bonferroni method
to adjust p-values.
Results: From 199 referrals, 36 GP care and 31 lifestyle arm patients completed to final follow-up. Significantly lower lifestyle arm
BAI scores at 20 weeks ( p b 0.001), non-significant ( p = 0.167) at 10 months after Bonferroni correction. 63.6% lifestyle arm, and
40% GP arm patients ( p = 0.045) panic-free at 20 weeks; 67.7% and 48.5% ( p = 0.123) respectively at 10 months.
Limitations: Final study size/power calls for caution in interpreting findings.
Conclusions: A lifestyle approach may provide a clinically effective intervention at least as effective as routine GP care, with significant
improvements in anxiety compared with routine GP care at the end of treatment. Further study is required before suggesting practice changes.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Clinical trial; Panic disorder; Lifestyle; Primary care; Occupational therapy; Pragmatic

☆
The main results from this research were presented at the 33rd
North American Primary Care Research Group Meeting held in
Quebec City on October 15th–18th 2005.
⁎ Corresponding author. Tel.: +44 1603 593096; fax: +44 1603 593166.
E-mail address: [email protected] (R.A. Lambert).
1
Clinical Trial Registration details: Controlled-trials.com—
ISRCTN51562655 http://www.controlled-trials.com/isrctn/search.asp.

0165-0327/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jad.2006.08.026

1. Introduction
In England, the cost of mental illness is estimated at
over £77 billion, including the costs of care provided by
the NHS, local authorities, privately funded services,
family and friends; lost output in the economy caused by

64

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

people being unable to work (paid and unpaid) and the
human costs of reduced quality of life, and loss of life,
amongst those experiencing a mental health problem (The
Sainsbury Centre for Mental Health, 2003).
Within the UK between 1994 and 1998, the prevalence of treated anxiety increased by over 30%, with
prevalence in females being consistently over twice that
of males. About 25% of GP consultations are for mental
health problems and 90% of mental health care is provided in primary care (Office of National Statistics,
2001). Although mental health problems are commonly
reported in primary care, these disorders are “often not
detected” (WHO, 2001 p. 59). When detected, current
clinical guidelines recommend either medication or psychological therapies (alone or in combination) (NICE,
2004). However, evidence of effective outcomes from
treatment is sometimes modest, or shows only small
effect sizes (Nadiga et al., 2003). A proportion of people
with mental health problems therefore continue to
experience morbidity, take time off work, and experience
social, psychological and physical consequences (WHO,
2001).
Panic attacks are a specific and severe form of anxiety
disorder, typified by the sudden onset of overwhelming
anxiety that presents with a variety of physical symptoms
such as palpitations, shortness of breath, dizziness and
nausea, and may involve fears of ‘going crazy’ or of
impending doom or death (Ohman, 2000). The prevalence of panic attacks more than doubled in the population
of the United States from 5.3% in 1980, to 12.7% in 1995
(Goodwin, 2003). Panic attacks occur in many anxiety
disorders and may be associated with specific events or
situations. However, panic attacks as a central feature of
panic disorder (PD) generally occur ‘out-of-the-blue’
(American Psychiatric Association, 2000). The prevalence of PD appears to be relatively consistent across
cultures at between 1% and 3% (Weissman et al., 1997).
Current treatment guidelines for PD (NICE, 2004),
recommend medication or psychological treatments such
as cognitive behavioural therapy (CBT) or counselling,
either individually, or in combination, as being the best
evidenced treatment. However, after 8 years of follow-up,
relapse rates of 64% in women and 21% in men were
reported in a study of 558 panic disorder patients enrolled
in the Harvard–Brown Anxiety Research Programme
(Yonkers et al., 2003).
A potential role for lifestyle in the prevention of
mental health problems has been recognised which relates to improving nutrition (Benton and Nabb, 2003;
Papakostas et al., 2004) and fluid intake (Muller et al.,
2002); reducing habitual substance use such as caffeine
(Alsene et al., 2003), alcohol (Davids et al., 2002) and

nicotine (Isensee et al., 2003); increasing exercise
(Callaghan, 2004) and reducing work stress (WHO,
2004). Nonetheless, very little mention is made of this
therapeutic approach within the current clinical guidance. The National Service Framework for Mental
Health (NSFMH) (DoH, 1999) recommends that modifying lifestyle factors such as smoking, excess alcohol
consumption; drug use; exercise and general physical
health may have a beneficial impact on mental health.
The present study was undertaken to test the hypothesis that modifying habitual lifestyle behaviours in a
health protective direction will reduce the frequency and
severity of anxiety and panic symptoms in panic disorder patients compared with usual GP care. It is
intended that a companion paper will report costeffectiveness findings.
2. Methods
An unblinded, pragmatic randomised controlled trial
was undertaken to determine whether an occupational
therapy-led lifestyle approach to the treatment of panic
disorder presenting in primary care achieves comparable
or better clinical outcomes when compared with routine
GP care.
2.1. Participants
Participants were adults (18–65 years) identified
with panic disorder with or without agoraphobia, from
15 GP practices within 2 Primary Care Trusts in the East
of England. At a screening interview with a trained
research therapist, eligibility was established through
administration of the Anxiety Disorders Interview
Schedule for DSM-IV (ADIS-IV). Participants were
required to meet the DSM-IV criteria for panic disorder
with or without agoraphobia. Participants prescribed
medication were required to have been taking a stable
dosage for at least 4 weeks before being entered into the
study. Results from the study show that 4 medication
types were commonly prescribed, and comprised over
75% of all prescriptions made during the trial. These
were SSRI's (43.5%), Benzodiazepines (15.2%), Betablockers (11.1%) and Tricyclics (5.8%). After providing
participants with a complete description of the study,
their written informed consent was obtained.
2.1.1. Stratification and randomisation
Participants were stratified in relation to the presence/
absence of agoraphobia and lifestyle factors to ensure
equal distribution of these characteristics within the
study arms. Agoraphobia was identified through the use

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

of the ADIS-IV (Brown et al., 1994), and lifestyle
factors were identified at initial interview by recording
average daily consumption of 1 or more of the
following: drinking more than 2 mugs of regular coffee;
drinking more than 3 mugs of tea; drinking more than 2
alcoholic drinks; drinking less than a total of 1.5 l of
fluid; smoking more than 5 cigarettes. After stratification, patients were randomly assigned to either routine
GP care or to the occupational therapy-led lifestyle
approach. Random allocation was carried out using
sealed envelopes prepared in advance by the principal
researcher and supplied to the research therapists.
2.1.2. Sample size
A power calculation was carried out using a standardised difference of 0.5 derived from a trial comparing
cognitive behavioural therapy, applied relaxation and
imipramine in the treatment of PD (Clark et al., 1994),
and a significance level of p = 0.05. For 90% power to
detect a 5 point difference on the BAI, a sample size of
170 patients (85 patients per arm) was estimated.
It was assumed that 50% of patients referred to the
study would either not meet the eligibility criteria, or not
consent to participate, and that 30% loss to follow-up
would occur. Just under 500 referrals into the study were
therefore estimated.
2.2. Design
Routine GP care was not defined, but was
monitored using routine GP practice records of
attendance, prescriptions and referrals. The trial
intervention addressed four areas of lifestyle; diet,
fluid intake, exercise and habitual lifestyle drug use
(caffeine, alcohol and nicotine). It provided up to 10
intervention sessions over a 16-week period (3 × 1 h
appointments at weekly intervals, 3 × half hour appointments at weekly intervals, 3 × half hour appointments at
fortnightly intervals and 1 × 1 h appointment at a
monthly interval). The intervention was delivered in
four stages:
• lifestyle review using self-report mood and lifestyle
diaries
• education to increase patient awareness of the
potential negative health effects of some lifestyle
behaviours (such as smoking and poor diet pattern)
and the health benefits of other lifestyle behaviours
(such as sufficient exercise and sufficient fluid intake)
• specific lifestyle changes (in diet, fluid intake, exercise
or habitual lifestyle drug use) were negotiated between
the therapist and the patient.

65

• monitoring and review of the agreed lifestyle changes
and any subsequent change in symptomatology.
2.3. Process measure
A process measure was required to document the
degree of lifestyle change during the study. The Lifestyle
Behaviour Index (LBI) was developed in a Dutch study
(Mulder et al., 1998) of the stability of lifestyle
behaviour in 1400 men over a 4-year period, and
provides a valid measure of change in smoking, alcohol
consumption, physical exercise and dietary habits. The
LBI does not, however, assess changes in caffeine use or
general fluid intake. Additional items were therefore
devised using the same format as in the original
questionnaire to assess these variables. In the analysis,
scores from both the original questionnaire and the
extended questionnaire are used.
2.4. Primary outcome measure
The Beck Anxiety Inventory (BAI), a broad indicator
of anxiety severity, was used as the primary outcome
measure. The BAI is a 21-item self-completion measure
of established validity and reliability that has been
previously used in studies of PD (de Beurs et al., 1997;
Clark et al., 1994). The BAI has a range of 0–63, with
bands indicating minimal (0–7), mild (8–15), moderate
(16–25) and severe (26–63) anxiety (Beck and Steer,
1993).
2.5. Secondary outcome measures
Secondary outcome measures were: panic attack
frequency, severity and symptomatology (ADIS-IV);
comorbid depression (Beck Depression inventory II
(BDI-II) (Beck et al., 1996); phobic symptoms (Fear
Questionnaire) (Marks and Mathews, 1979); and quality
of life (Short Form Health Survey Questionnaire (SF36) (Ware, 1993) and EuroQol Health Questionnaire
(EQ-5D) (Kind and de Charro, 1998).
2.6. Analysis
While the intervention period was over 16 weeks,
analysis was carried out at baseline, and at follow-up
appointments at 20 weeks and at 10 months after randomisation. Each of the outcome measures was provided to
participants at baseline and at each follow-up by the
research therapists who were different from the therapists providing the intervention. These were also
collected at the interview and placed in a file ready for

66

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

data entry and analysis at the end of the trial. Analysis
consisted primarily of between-group comparisons using
Chi squared tests for categorical outcome measures;
unpaired t-tests (Mann–Whitney U-tests if data were not
normally distributed) and analysis of covariance
(ANCOVA) to adjust for minor baseline differences in
the degree of apprehension, degree of interference in
daily life and the number of limited symptom attacks,
along with stratification factors of agoraphobia and
lifestyle factors. Within group changes were analysed
using paired t-tests.
To allow for the effect of performing multiple tests of
significance, the Bonferroni method was employed to
adjust p-values (reducing the level at which results were
considered significant from p = 0.05 to p = 0.007) (Pocock
et al., 2002a). Analysis was carried out on an intention-totreat basis where all patients on whom information was
available were analysed within the groups to which they
were originally allocated (Petrie and Sabin, 2003).
3. Results
Despite initially positive indications for recruitment
rates from pilot data, the number of referrals received
between February 2001 and July 2003 was considerably
lower than required (199 patients) with an uneven
distribution across the 15 participating GP practices
(Mean referrals/practice = 13.3; Min = 1; Max = 32). From
these referrals, 59% (117 patients) were entered into the
study, and 43% (50 patients) were lost to follow-up, with
similar losses within each study arm. No indication was
received of the reasons for withdrawal from the study. The
median number of intervention sessions attended in the
lifestyle arm was 5 (range 0–10). The study retained 31
Lifestyle arm patients and 36 GP care arm patients to final
analysis (See Fig. 1). This resulted in a study which was
underpowered in comparison to that planned. Rather than

carry out a post-hoc power calculation which is viewed
with suspicion amongst some statisticians (Levine and
Ensom, 2001; Pocock et al., 2002c), ANCOVA was used
to report mean difference and 95% confidence intervals as
a more appropriate indicator of sample size issues.
In this study therefore, 67 patients completed to final
follow-up and 50 patients did not. Only minor differences
were observed between completers and non-completers in
mean age (completers 40.2; non-completers 38.2); gender
(completers 70% female; non-completers 66% female)
and mean number of children (completers 1.43; noncompleters 1.54). There were also no significant differences in any clinical measures. There were no adverse
events reported during the trial.
Clinical comparisons were made which did not
show any significant differences between these two
groups on any baseline measure. Patients completing
the research may be considered therefore as being
representative of the patient group who were eligible
for the study, and patients not completing the study
were unlikely to be significantly different from those
completing the study.
A comparison of baseline demographic and clinical
characteristics of patients in each study arm is provided
in Table 1, and show broadly comparable groups. In the
period of 3 months before entry into the study, 9
referrals (including 2 mental health referrals) were made
for specialist consultations outside the GP practice for
patients entered into the lifestyle arm, compared with 12
(8 mental health) for patients entered into the GP care
arm. Analysis of covariance was, however, used to
adjust for the small observed differences in the number
of limited symptom attacks (as a between-group
imbalance was observed at baseline), degree of
interference and degree of apprehension in daily life
(as predictors of outcome), the LBI extended total score
and the fear questionnaire total phobia score (as factors

Fig. 1. Study progression.

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

67

Table 1
Baseline demographic and clinical characteristics
Lifestyle arm
N
Gender (female)
Age
No. of children

No. of GP Practice appointments in last 3 months
Total lifestyle factors
ADIS-IV
Number of full (4 or more symptoms) panic attacks/last month
Number of limited symptom attacks (3 or less symptoms/last month)
Degree of apprehension
Degree of interference in daily life
Beck Anxiety Inventory
Beck Depression Inventory
Fear Questionnaire
Anxiety and depression score
Total score
SF-36 Health Survey
SF Physical Function
SF Role—Physical
SF Bodily Pain
SF General Health
SF Vitality
SF Social Functioning
SF Role—Emotional
SF Mental Health
EQ-5D EuroQol
Health state coding
Valuation of own health/100
Lifestyle Behaviour Index
Smoking
Alcohol
Physical exercise
Diet
Caffeine
General fluid
Original Total
Extended Total

used in stratifying the randomisation procedure)
(Pocock et al., 2002b).

Mean

Routine GP care arm
%

N

68.4

41/60
60
56

38.6
1.3

S.D.

N

Mean

3.0
2.0

2.7
1.0

46
59

2.8
1.5

3.0
1.0

1.9
0.9

7.4
15.6
5.5
5.9
29.5
24.3

5.0
8.0
6.0
6.0
29.0
23.5

6.4
19.5
2.1
1.9
11.3
10.5

59
56
59
58
58
58

8.5
8.7
5.4
5.1
29.4
23.1

5.0
6.0
6.0
5.0
27.0
23.0

9.0
10.3
2.3
1.7
13.5
12.3

54
54

21.1
61.4

22.0
59.5

8.2
23.1

57
57

21.8
61.8

22.0
62.0

8.1
24.9

54
54
53
54
54
53
53
54

82.2
60.2
41.6
53.3
32.8
59.7
25.2
42.2

90.0
75.0
42.0
52.0
30.0
62.5
0.0
42.0

19.8
40.8
10.1
15.4
19.0
24.2
33.9
15.1

59
59
59
57
59
59
59
59

80.5
59.3
43.0
54.5
36.5
60.8
36.2
45.3

90.0
75.0
44.0
57.0
35.0
62.5
33.3
48.0

23.6
41.5
10.0
17.6
20.6
28.9
40.7
17.7

53
53

0.6
60.8

0.7
60.0

0.3
18.8

59
58

0.7
64.2

0.8
67.5

0.2
18.1

52
54
54
54
53
54
52
52

4.1
2.1
1.5
1.6
1.0
2.4
9.3
12.8

2.0
2.0
1.0
2.0
1.0
2.0
9.0
13.0

4.1
1.7
1.2
1.3
0.9
1.3
51.0
5.0

58
58
58
58
58
58
58
58

4.0
2.2
1.6
1.7
0.8
2.0
9.4
12.3

1.0
2.0
1.0
2.0
1.0
2.0
8.0
10.5

43.0
1.7
1.5
1.5
0.7
1.4
6.3
7.1

39/57
57
55

40.1
1.7

N

Mean

39
54

3.4
1.7

55
52
55
55
54
54

Median

Mean

%
68.3

Median

S.D.

either follow-up for smoking or alcohol use. In contrast,
there were no significant changes in any LBI variable at
any trial stage in the GP care arm.

3.1. Between-group comparisons
3.1.1. Process measure (Lifestyle Behaviour Index (LBI))
Fig. 2 shows that significant LBI changes were observed in the lifestyle arm at both 20 weeks and at
10 months for physical exercise, diet, general fluid intake,
and both original and extended total scores. A significant
change in caffeine use between baseline and 20 weeks was
non-significant between baseline and 10 months. No
significant changes were observed between baseline and

3.1.2. Primary outcome measure (BAI)
At baseline, over 85% of patients in both study arms
were observed to have moderate to severe anxiety. In the
lifestyle arm, fewer than 20% of patients experienced
severe anxiety at 20 weeks, although at 10 months this
had risen to 30%. In the GP care arm, the level of severe
anxiety reduced to 55% at 20 weeks, and reduced further
to just over 40% at 10 months. ANCOVA (see Fig. 3)
shows results were significantly different at 20 weeks

68

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

Fig. 2. Lifestyle arm paired t-test for lifestyle behaviour index (Baseline—assessment period).

(after Bonferroni correction) for the total BAI score
( p b 0.001, mean difference (md) −9.8; 95% confidence
interval (CI) − 15.0 to − 4.6); the neurophysiological
( p = 0.002), subjective ( p b 0.001) and autonomic
( p = 0.006) subscores, but not for the panic subscore
( p = 0.041). There were no significant between-group
differences at 10 months.
A proportion of patients (36% lifestyle arm; 50% GP
care arm) had higher levels of anxiety at 10 months than
at 20 weeks. When mean BAI scores at 10 months were
compared with those at baseline, a significantly greater
number (Fishers Exact Test (2-sided) p = 0.016) of
lifestyle arm patients remained improved (28/29
patients) than GP care arm patients (25/34 patients).
This is illustrated by the finding that at 10 months just
over 26% (9/34) of GP care arm patients compared with
3.4% (1/29) lifestyle arm patients were observed to have
an increased BAI score.

3.1.3. Secondary outcome measures
At 20 weeks, 63.6% of lifestyle arm patients had
been panic-free in the previous month compared with
40% in the GP care arm. At the 10-month follow-up,
67.7% of lifestyle arm patients and 48.5% of GP care
arm patients had been panic-free for 1 month. Although
there was therefore an observed trend towards greater
improvement in panic reactions in the lifestyle arm, the
differences were not statistically significant.
In the period of 3 months before follow-up at
20 weeks, 4 further referrals (2 mental health) had been
made for lifestyle arm patients and 6 further referrals (3
mental health) for GP care arm patients. In the 3-month
period before follow-up at 10 months, 6 further lifestyle
arm referrals (0 mental health) were made, compared
with 17 (7 mental health) referrals for GP care arm
patients.
A range of other secondary outcome measures showed
significant or near significant results at 20 weeks, but only
the general health item from the SF36 showed a significant result at 10 months (before Bonferroni correction). After Bonferroni correction, only the anxiety and
depression score of the fear questionnaire at 20 weeks
remained significant. Once again this indicates a trend
towards improvement in the lifestyle arm, but not at a
statistically significant level.
4. Discussion

Fig. 3. ANCOVA for BAI results at 20 weeks and 10 months.

The main conclusion from this study is that there is
some evidence that a lifestyle approach in the treatment

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

of PD in primary care may be at least as clinically
effective as routine GP care, and may be more effective
over a short follow-up period.
At baseline, patients in both the lifestyle and routine
GP care arms were broadly demographically and
clinically similar. The pattern of results at both
20 weeks and at 10 months consistently favoured the
lifestyle arm. At 20 weeks, differences favouring the
lifestyle intervention compared with routine GP care
remained significant after Bonferroni correction on the
BAI total score, ( p b 0.001), and on all except the BAI
panic subscores. At the 10-month follow-up however, no
significant between-group differences were observed.
The mean BAI score at 10 months improved by 5 or more
points in over 85% (25/29) of lifestyle patients and 60%
(21/34) of GP care arm patients. At 10 months, over 26%
(9/34) of GP care arm patients were observed to have an
increased BAI total score compared with 3.4% (1/29) of
patients in the lifestyle arm. These findings indicate that
while there were only minor between-group differences in
mean BAI scores at the 10 month follow-up, at 20 weeks a
significant short-term benefit was observed. The lifestyle
approach also appears to have a significantly lower
relapse rate than routine GP care.
One possible factor in the lifestyle arm relapse rate
after the end of treatment at 20 weeks may be the duration
of lifestyle treatment. The possibility of using maintenance appointments (possibly at 2 to 3 monthly intervals
over the 12 months following the end of treatment) to
improve clinical outcomes therefore requires further examination. It should also be noted that in a comparison of
study effects (see Fig. 4) from studies using the BAI as an

69

outcome measure, the degree of improvement from
baseline to the end of treatment and at follow-up in the
lifestyle arm compares favourably with CBT, other psychological treatments, Imipramine (Clark et al., 1994,
1999; Proudfoot et al., 2003a), and treatment as usual
(Proudfoot et al., 2003b).
Only relatively minor between-group differences
were observed using the secondary outcome measures.
From 199 referrals received, a higher than expected
percentage of patients were randomised (59%), but there
was also a higher than expected loss to follow-up (43%).
The study therefore fell short of the required sample
size based on 90% power, but nonetheless there is a
significant difference after Bonferroni correction on the
primary outcome measure at first follow-up. The aim of
treatment in the lifestyle arm was the promotion of
positive lifestyle changes. Significant LBI changes were
observed in the lifestyle arm on all except smoking and
alcohol subscores. No such changes were observed in
the GP care arm. The lifestyle intervention therefore
appears to promote lifestyle changes across a range of
habitual behaviours, and these appear to have been
maintained at 10 months in about two-thirds of lifestyle
arm patients. However, smoking and alcohol consumption were more resistant to change.
A range of possible mechanisms may underlie the
clinical findings. About 40% of PD is genetically
determined, and this is expressed in some PD patients
through altered sensitivity in a range of normal body
functions compared with other patient groups and
normal controls. Such altered sensitivity in PD patients
has been reported in the respiratory (Horwath et al.,

Fig. 4. Difference in BAI total score between baseline and end of treatment and baseline and follow-up.

70

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71

1997; Biber and Alkin, 1999), cardiovascular (Yeragani
et al., 2003), musculoskeletal (Gulpek et al., 2004),
audio-vestibular (Perna et al., 2001; Asmundson et al.,
1998), gastro-intestinal (Kumano et al., 2004; Maunder,
1998) and immune systems (Slattery et al., 2002; Koh
and Lee, 2004). A combination of habitual lifestyle
behaviours such as diet, fluid intake exercise and
habitual lifestyle drug use (such as caffeine, alcohol
and nicotine) may interact with altered sensitivity in
body system function in some PD patients.
It is proposed that the interaction between altered
body system sensitivity, and adverse habitual lifestyle
behaviours in PD patients may provoke cognitive
awareness of altered bodily sensations at an earlier
stage than for other groups. Cognitive appraisal of these
physiological symptoms may then provoke further
physiological and cognitive reactions, thereby raising
anxiety levels into an anxiety spiral leading to a panic
attack. The latter mechanism is similar to the original
cognitive model proposed by Clarke (1986). It deviates
from this model in the proposal of a physiological
trigger, possibly resulting from a combination of lowlevel symptoms from a range of body systems that reach
an additive threshold level before reaching cognitive
awareness.
A review of habitual lifestyle behaviours can identify
possible triggers for physiological reactions underlying
panic responses. Using lifestyle review and lifestyle
change in this way can provide PD patients with a
rational explanation of their symptoms, a range of
possible strategies through which to explain and alter the
sensations they are experiencing, and thus placing
symptoms under their own control. Although the
intervention in the present study was provided by
occupational therapists, the efficacy of using other
professionals such as Practice Nurses, or training Health
Care Assistants in the provision of this form of intervention should be examined.
In conclusion, lower than expected numbers of
patients in the study suggest caution should be exercised
in interpreting results. However, significant results after
Bonferroni correction and reporting of point estimates
and 95% confidence intervals show that an occupational
therapy-led lifestyle approach to the treatment of PD in
primary care represents a clinically effective treatment at
least equal to that of routine GP care. At the end of
treatment, a significantly greater number of patients
receiving the lifestyle intervention recorded lower levels
of anxiety than at baseline, compared with those
receiving routine GP care.
By the 10-month follow-up, while the results generally
favoured the lifestyle intervention, the between-group

differences were statistically non-significant. The results
from the present study show that this approach does not
work in the longer term, for all patients, as a return to
previous lifestyle and anxiety patterns was observed in
about one-third of patients. However, in the remaining
two-thirds of patients in the lifestyle arm of the trial,
positive lifestyle changes were promoted and were
maintained at the 10-month follow-up, and were
associated with continued improvements in anxiety.
Acknowledgements
To the grant awarding bodies whose financial support
made the research possible. To the patients, the GP
Practices and Occupational Therapists who participated
in the research, without whose continued support the
research could never have been achieved.
The Principal Investigator and corresponding author
who takes responsibility for the integrity of the data and
the accuracy of the data analysis, and that all authors had
full access to all the data in the study.
This research was supported by a NHS R&D National
Primary Care Researcher Development Award Fellowship (Grant RDA99/062) and by a NHS R&D Eastern
Region Health Services and Public Health Research
Scheme Grant (Grant HSR/0500/1).
References
Alsene, K., Deckert, J., Sand, P., de Wit, H., 2003. Association
between A2a receptor gene polymorphisms and caffeine-induced
anxiety. Neuropsychopharmacology 28, 1694–1702.
American Psychiatric Association, 2000. Diagnostic and Statistical
Manual of Mental Disorders; DSM-IV TR. American Psychiatric
Association, Washington, DC.
Asmundson, G.J., Larsen, D.K., Stein, M.B., 1998. Panic disorder and
vestibular disturbance: an overview of empirical findings and
clinical implications. J. Psychosom. Res. 44, 107–120.
Beck, A.T., Steer, R.A., 1993. Beck Anxiety Inventory Manual. The
Psychological Corporation, San Antonio.
Beck, Aaron T., Steer, Robert A., Brown, Gregory K. Beck Depression
Inventory-BDI-II Manual. 1996. San Antonio, The Psychological
Corporation, Harcourt Brace and Co.
Benton, D., Nabb, S., 2003. Carbohydrate, memory, and mood. Nutr.
Rev. 61, S61–S67.
Biber, B., Alkin, T., 1999. Panic disorder subtypes: differential
responses to CO2 challenge. Am. J. Psychiatry 156, 739–744.
Brown, T.A., DiNardo, P.A., Barlow, D.H., 1994. Anxiety Disorders
Interview Schedule for DSM-IV (Adult version). Graywind
Publications Incorporated.
Callaghan, P., 2004. Exercise: a neglected intervention in mental health
care? J. Psychiatr. Ment. Health Nurs. 11, 476–483.
Clark, D.M., 1986. A cognitive approach to panic. Behav. Res. Ther.
24, 461–470.
Clark, D.M., Salkovskis, P.M., Hackmann, A., Middleton, H.,
Anastasiades, P., Gelder, M., 1994. A comparison of cognitive

R.A. Lambert et al. / Journal of Affective Disorders 99 (2007) 63–71
therapy, applied relaxation and imipramine in the treatment of
panic disorder [see comments]. Br. J. Psychiatry 164, 759–769.
Clark, D.M., Salkovskis, P.M., Hackmann, A., Wells, A., Ludgate, J.,
Gelder, M., 1999. Brief cognitive therapy for panic disorder: a
randomized controlled trial. J. Consult Clin. Psychol. 67, 583–589.
Davids, E., Muller, M.J., Rollmann, N., Burkart, M., Regier-Klein, E.,
Szegedi, A., Benkert, O., Maier, W., 2002. Syndrome profiles in
alcoholism and panic disorder with or without agoraphobia: an
explorative family study. Prog. Neuro-psychopharmacol. Biol.
Psychiatry 26, 1079–1087.
de Beurs, E., Wilson, K.A., Chambless, D.L., Goldstein, A.J., Feske,
U., 1997. Convergent and divergent validity of the Beck Anxiety
Inventory for patients with panic disorder and agoraphobia.
Depress. Anxiety 6, 140–146.
DoH, 1999. National Service Framework for Mental Health: Modern
Standards and Service Models. Department of Health, London.
Goodwin, R.D., 2003. The prevalence of panic attacks in the United
States: 1980 to 1995. J. Clin. Epidemiol. 56, 914–916.
Gulpek, D., Bayraktar, E., Akbay, S.P., Capaci, K., Kayikcioglu, M.,
Aliyev, E., Soydas, C., 2004. Joint hypermobility syndrome and
mitral valve prolapse in panic disorder. Prog. Neuro-psychopharmacol. Biol Psychiatry 28, 969–973.
Horwath, E., Adams, P., Wickramaratne, P., Pine, D., Weissman, M.M.,
1997. Panic disorder with smothering symptoms: evidence for
increased risk in first-degree relatives. Depress. Anxiety 6,
147–153.
Isensee, B., Wittchen, H.U., Stein, M.B., Hofler, M., Lieb, R., 2003.
Smoking increases the risk of panic: findings from a prospective
community study. Arch. Gen. Psychiatry 60, 692–700.
Kind, Paul, de Charro, Frank, 1998. Euroqol EQ-5D: User guide.
Centre for Health Economics, University of York, York.
Koh, K.B., Lee, Y., 2004. Reduced anxiety level by therapeutic
interventions and cell-mediated immunity in panic disorder
patients. Psychother. Psychosom. 73, 286–292.
Kumano, H., Kaiya, H., Yoshiuchi, K., Yamanaka, G., Sasaki, T.,
Kuboki, T., 2004. Comorbidity of irritable bowel syndrome, panic
disorder, and agoraphobia in a Japanese representative sample.
Am. J. Gastroenterol. 99, 370–376.
Levine, M., Ensom, M.H., 2001. Post hoc power analysis: an idea
whose time has passed? [see comment]. Pharmacotherapy 21,
405–409.
Marks, I.M., Mathews, A.M., 1979. Brief standard self-rating for
phobic patients. Behav. Res. Ther. 17, 263–267.
Maunder, R.G., 1998. Panic disorder associated with gastrointestinal
disease: review and hypotheses. J. Psychosom. Res. 44, 91–105.
Mulder, M., Ranchor, A.V., Sanderman, R., Bouma, J., van den
Heuvel, W.J., 1998. The stability of lifestyle behaviour. Int. J.
Epidemiol. 27, 199–207.
Muller, V., Birbaumer, N., Preissl, H., Braun, C., Lang, F., 2002.
Effects of water on cortical excitability in humans. Eur. J.
Neurosci. 15, 528–538.
Nadiga, D.N., Hensley, P.L., Uhlenhuth, E.H., 2003. Review of the
long-term effectiveness of cognitive behavioral therapy compared
to medications in panic disorder. Depress. Anxiety 17, 58–64.
NICE, 2004. Clinical Guidelines for the Management of Panic
Disorder and Generalised Anxiety Disorder. University of
Sheffield, ScHARR.
Office of National Statistics, 2001. Prevalence of Treated Anxiety per
1000 Patients, by Age, Sex and Calendar Year: 1994–98. Office of
National Statistics, London.

71

Ohman, A., 2000. Anxiety. In: Fink, G. (Ed.), Encyclopedia of Stress.
Academic Press, London, pp. 226–231.
Papakostas, G.I., Ongur, D., Iosifescu, D.V., Mischoulon, D., Fava,
M., 2004. Cholesterol in mood and anxiety disorders: review of the
literature and new hypotheses. Eur. Neuropsychopharmacol. 14,
135–142.
Perna, G., Dario, A., Caldirola, D., Stefania, B., Cesarani, A., Bellodi,
L., 2001. Panic disorder: the role of the balance system.
J. Psychiatr. Res. 35, 279–286.
Petrie, A., Sabin, C., 2003. Medical Statistics at a Glance. Blackwell
Science, Padstow.
Pocock, S.J., Assmann, S.E., Enos, L.E., Kasten, L.E., 2002a. Subgroup
analysis, covariate adjustment and baseline comparisons in clinical
trial reporting: current practice and problems. Stat. Med. 21,
2917–2930.
Pocock, S.J., Assmann, S.E., Enos, L.E., Kasten, L.E., 2002b. Subgroup
analysis, covariate adjustment and baseline comparisons in clinical
trial reporting: current practice and problems. Stat. Med. 21,
2917–2930.
Pocock, S.J., Assmann, S.E., Enos, L.E., Kasten, L.E., 2002c. Subgroup
analysis, covariate adjustment and baseline comparisons in clinical
trial reporting: current practice and problems. Stat. Med. 21,
2917–2930.
Proudfoot, J., Goldberg, D., Mann, A., Everitt, B., Marks, I., Gray, J.A.,
2003a. Computerized, interactive, multimedia cognitive–behavioural program for anxiety and depression in general practice.
Psychol. Med. 33, 217–227.
Proudfoot, J., Goldberg, D., Mann, A., Everitt, B., Marks, I., Gray, J.A.,
2003b. Computerized, interactive, multimedia cognitive–behavioural program for anxiety and depression in general practice.
Psychol. Med. 33, 217–227.
Slattery, M.J., Klein, D.F., Mannuzza, S., Moulton III, J.L., Pine, D.S.,
Klein, R.G., 2002. Relationship between separation anxiety
disorder, parental panic disorder, and atopic disorders in children:
a controlled high-risk study. J. Am. Acad. Child Adolesc. Psych.
41, 947–954.
The Sainsbury Centre for Mental Health, 2003. Economic and Social
Costs of Mental Illness in England. The Sainsbury Centre for
Mental Health, London.
Ware, J.E., 1993. SF-36 Health Survey: Manual and Interpretation Guide.
The Health Institute, New England Medical Centre, Massachusetts.
Weissman, M.M., Bland, R.C., Canino, G.J., Faravelli, C., Greenwald,
S., Hwu, H.G., Joyce, P.R., Karam, E.G., Lee, C.K., Lellouch, J.,
Lepine, J.P., Newman, S.C., Oakley-Browne, M.A., Rubio-Stipec,
M., Wells, J.E., Wickramaratne, P.J., Wittchen, H.U., Yeh, E.K.,
1997. The cross-national epidemiology of panic disorder. Arch.
Gen. Psychiatry 54, 305–309.
WHO, 2001. World Health Report 2001: Mental Health, New
Understanding, New Hope. World Health Organisation, Geneva.
WHO, 2004. Prevention of Mental Disorders: Effective Interventions
and Policy Options. World Health Organisation, Geneva.
Yeragani, V.K., Tancer, M., Uhde, T., 2003. Heart rate and QT interval
variability: abnormal alpha-2 adrenergic function in patients with
panic disorder. Psychiatry Res. 121, 185–196.
Yonkers, K.A., Bruce, S.E., Dyck, I.R., Keller, M.B., 2003.
Chronicity, relapse, and illness-course of panic disorder, social
phobia, and generalized anxiety disorder: Findings in men and
women from 8 years of follow-up. Depress. Anxiety 17, 173–179.