Cost-effectiveness of psychological and pharmacological interventions for generalized anxiety disorder and panic disorder

Louise Heuzenroeder, Marie Donnelly, Michelle M. Haby,

Cathrine Mihalopoulos, Ruth Rossell, Rob Carter, Gavin Andrews, Theo Vos

Objective:

To assess from a health sector perspective the incremental cost-effectiveness ofinterventions for generalized anxiety disorder (cognitive behavioural therapy [CBT] andserotonin and noradrenaline reuptake inhibitors [SNRIs]) and panic disorder (CBT, selectiveserotonin reuptake inhibitors [SSRIs] and tricyclic antidepressants [TCAs]).

Method:

The health benefit is measured as a reduction in disability-adjusted life years(DALYs), based on effect size calculations from meta-analyses of randomised controlledtrials. An assessment on second stage filters (‘equity’, ‘strength of evidence’, ‘feasibility’ and‘acceptability to stakeholders’) is also undertaken to incorporate additional factors that impacton resource allocation decisions. Costs and benefits are calculated for a period of one yearfor the eligible population (prevalent cases of generalized anxiety disorder/panic disorderidentified in the National Survey of Mental Health and Wellbeing, extrapolated to theAustralian population in the year 2000 for those aged 18 years and older). Simulationmodelling techniques are used to present 95% uncertainty intervals (UI) around theincremental cost-effectiveness ratios (ICERs).

Results:

Compared to current practice, CBT by a psychologist on a public salary is the mostcost-effective intervention for both generalized anxiety disorder (A$6900/DALY saved; 95%UI A$4000 to A$12 000) and panic disorder (A$6800/DALY saved; 95% UI A$2900 toA$15 000). Cognitive behavioural therapy results in a greater total health benefit than thedrug interventions for both anxiety disorders, although equity and feasibility concerns for CBTinterventions are also greater.

Conclusions:

Cognitive behavioural therapy is the most effective and cost-effective inter-vention for generalized anxiety disorder and panic disorder. However, its implementationwould require policy change to enable more widespread access to a sufficient number oftrained therapists for the treatment of anxiety disorders.

Key words:

antidepressive agents, anxiety disorders, cognitive behaviour therapy, cost

Australian and New Zealand Journal of Psychiatry 2004; 38:602–612

effectiveness, meta-analysis.

Marie Donnelly, Project Officer; Michelle Haby, Epidemiologist (Corres-pondence); Theo Vos, Senior Epidemiologist

Health Surveillance and Evaluation Section, Public Health, Departmentof Human Services, Melbourne, Victoria, Australia. Email: michelle.haby@dhs.vic.gov.au

Louise Heuzenroeder, Research Fellow

Victorian Public Health Training Scheme, Department of HumanServices, Melbourne, Victoria, Australia

Cathrine Mihalopoulos, Research Fellow; Rob Carter, Deputy Director

Program Evaluation Unit, University of Melbourne, Melbourne, Vic-toria, Australia

Ruth Rossell, Researcher; Gavin Andrews, Director

CRUfAD, University of New South Wales, Sydney, New South Wales,Australia

Received 29 August 2003; revised 20 April 2004; accepted 21 April 2004.

L. HEUZENROEDER, M. DONNELLY, M.M. HABY, C. MIHALOPOULOS, R. ROSSELL, R. CARTER, G. ANDREWS, T. VOS 603

Generalized anxiety disorder (GAD) contributes sig-nificantly to the burden of disease, ranking twelfth inwomen and twenty-fourth in men as a cause of morbidityin Australia [1]. Panic disorder (PD) is less common thanGAD, but the most common diagnosis among peopleseeking treatment for an anxiety problem [2]. Using theICD-10 definitions [3], the National Survey of MentalHealth and Wellbeing (NSMHW) [4] gives a 12-monthprevalence estimate of 3.0% for GAD and 1.4% forpanic disorder (defined as panic disorder with or withoutagoraphobia).

Treatment guidelines [5] suggest the primary treat-ments for GAD should be non-pharmacological. Wherepharmacological intervention is required, they recom-mend the antidepressant venlafaxine (a serotonin andnoradrenaline reuptake inhibitor [SNRI]) or buspirone.However, buspirone is rarely prescribed in Australia.Benzodiazepines are recommended for acute exacerba-tions [5], but are not considered here as there is inade-quate evidence to support long-term use. However, notethat the national BEACH program [6] tells us that 44%of GP encounters for ‘anxiety’ result in a prescription forbenzodiazepines (additional analysis of BEACH data:April 2000 to March 2001).

While the guidelines specifically recommend venla-faxine, the Cochrane review ‘Antidepressants for GAD’found that venlafaxine, paroxetine (a selective serotoninreuptake inhibitor [SSRI]) and imipramine (a tricyclicantidepressant [TCA]) have comparable efficacy com-pared to placebo [7]. The Cochrane review measuresefficacy using dichotomous outcome measures. Thecurrent analysis relies on continuous outcome measuresreporting means and standard deviations to calculateefficacy. Therefore, many RCTs referred to in theCochrane review are excluded [8–14]. Neither paroxet-ine nor venlafaxine are listed on the PharmaceuticalBenefits Scheme (PBS) for GAD [15]. Imipramine islisted without restrictions.

For PD, clinical practice guidelines recommend cogni-tive behavioural therapy (CBT) as the most consistentlyefficacious treatment and the treatment that should beattempted initially [5]. Some people with panic disordermay require additional pharmacological treatment, ormay not respond to CBT. In this case, the TherapeuticGuidelines [5] recommend SSRIs, and then benzodi-azepines, TCAs and irreversible non-selective monoamineoxidase inhibitors (MAOIs). By contrast, guidelines bythe Royal Australian College of Physicians recommendTCAs before benzodiazapines and SSRIs [16], notingthat medication with benzodiazapines may result independence and long-term adverse effects, and that evi-dence for treating panic with MAOIs is currently insuf-ficient. Paroxetine is the only SSRI listed on the PBS for

PD, and imipramine is the only TCA that is listedwithout restrictions for PD [15]. The interventionsanalyzed for PD in this report are CBT, SSRIs (parox-etine) and TCAs (imipramine).

The aim of this study is to assess the incremental cost-effectiveness of CBT and SNRIs for the treatment ofGAD and CBT, SSRIs and TCAs for PD, for those aged18 years and older. The analysis of interventions forGAD and PD contribute to the state and Commonwealthgovernment-funded Assessing Cost-Effectiveness – MentalHealth (ACE-MH) project, in which economic evalua-tions are performed for a range of interventions formental disorders, using common methods [17].

Method

The ICER is calculated as the cost (A$) per disability-adjusted lifeyear (DALY) saved. The population eligible for the interventions areadults (aged 18 years and older) with ICD-10 defined GAD/PD in theAustralian population in the year 2000 who sought health care for theirmental health disorder but would not have received evidence-basedmedicine (EBM) under current practice. All interventions involve theprovision of treatments for which there is evidence of efficacy.

These analyses are performed from the perspective of the health caresector; therefore, both government and patient expenses for servicesand pharmaceuticals are included. As the tracking of costs and benefitsis for one year, discounting is not applied.

The interventions

Cognitive behavioural therapy

(for GAD and PD): The CBT inter-ventions are modelled as 12 one-hour consultations plus one generalpractitioner (GP) consultation for referral to one of four differentproviders: psychologists and psychiatrists, either working privately oron a publicly funded salary.

Serotonin and noradrenaline reuptake inhibitors – venlafaxine

(for GAD): The SNRI intervention modelled is 12 months of venla-faxine (at 75 mg or 150 mg per day) plus doctor consultations. Theinclusion of both doses reflects current practice, based on data fromthe pharmaceutical benefits scheme. We assume the effect size foreach dose is the same. This is because our meta-analysis of two RCTs[18,19] revealed similar effect sizes (0.42 for 75 mg and 0.46 for150 mg) and overlapping confidence intervals for each dose; hence,outcomes of each dose from both trials were combined to produce asingle effect size of 0.44 (95% CI = 0.28–0.60).

The number of doctor consultations is assumed to be nine per yearfor all drug interventions: weekly visits for the first month, monthlyvisits for the next two months and three-monthly visits thereafter. Wederive the proportion of consumers consulting with a GP or referred toa psychiatrist for treatment from the NSMHW, assuming this reflectscurrent practice (approximately 10% referred to a psychiatrist, 90%managed by a GP [4]). For those consulting a psychiatrist, a GP visitfor referral is also included.

Serotonin reuptake inhibitors – paroxetine

(for PD): The dosemodelled is 40 mg per day, as recommended in the therapeutic guide-lines; it is also the average dose in the RCTs on which effect size and

604 INTERVENTIONS FOR ANXIETY DISORDERS

adherence are based. In PD approximately 26% of patients are referredto a psychiatrist and 74% are managed by GPs.

Tricyclic antidepressants – imipramine

(for PD): A dose of225 mg per day of imipramine is modelled – the average dose used inthe RCTs on which the effect size is based.

Current practice

Current practice for the treatment of GAD and PD is determinedfrom the NSMHW by identifying service utilization patterns for thosewith the disorder who consulted health services for a mental healthproblem. All prevalent cases of GAD and PD in the year 2000 areincluded. Consulting is defined as seeking care for a mental healthproblem during the past 12 months from a general practitioner, psychi-atrist, psychologist, physician, surgeon, social worker, mental healthteam worker or an admission to hospital [4]. It is assumed that thoseconsulting are receiving EBM if they have had three or more consulta-tions with a GP, psychiatrist or psychologist plus CBT and/or drugtreatment.

Of those with GAD, 55% consulted; these were further split intotreatment with EBM (27%) or non-EBM (28%). Those who consultedand received non-EBM under current practice each averaged 2.5 GPconsultations, 0.2 psychiatrist consultations and 0.6 psychologist con-sultations within the past 12 months. For PD, 65% consulted, with 47%receiving EBM and 19% receiving non-EBM. Those who consultedand received non-EBM under current practice each averaged 2.8 GPconsultations, 0.3 psychiatrist consultations and 1.8 psychologist con-sultations within the past 12 months.

Assessment of benefit

Benefits are assessed by a two-stage process. In the first stage,DALYs are employed to estimate the health benefit from the interven-tions. The second stage involves the assessment of issues that eitherinfluence the degree of confidence that can be placed in the ICERs(such as the level of available evidence), or broader issues that need tobe taken into account in decision-making about resource allocation(such as equity and acceptability to stakeholders).

Stage one: measurement of the health gain

The health benefit is measured in DALYs. There is no evidence inthe literature that SNRIs, SSRIs, TCAs or CBT can cause or preventdeath, so only a change in the years lived with disability (YLD) com-ponent of the DALY is modelled.

YLD for the current practice comparator

The one-year prevalence estimates from the NSMHW [4,20] are437 417 prevalent cases of GAD and 193 641 prevalent cases of PD inAustralia (Table 1).

Because GAD and panic disorder are chronic conditions withperiods of remission and relapse evident for up to 20 years [21–23], theinterventions will not be producing health gain at all times. Therefore,the health gain is multiplied by the proportion of time symptomatic,which is estimated from the NSMHW as the current prevalence (symp-tomatic in the past 2 weeks for GAD, and past 4 weeks for PD) divided

by the prevalence in the past 12 months. The percentage of timesymptomatic over 12 months is estimated to be 63% for GAD and 48%for PD.

The disability weights (DWs) used to calculate YLDs are based on theDutch weights. The DWs for GAD are 0.17 for mild and moderate cases,and 0.60 for severe cases. Those for PD are 0.16 for mild and moderate,and 0.69 for severe [24]. Composite DWs were calculated separately forthose who did not consult, those who consulted and received EBM, andthose who consulted and received non-EBM. We determine the spreadof severity of the disorders using the NSMHW, assuming this reflects thespread of severity in the Australian population. Severity is classifiedusing the Mental Component Score (MCS) of the SF-12, which has amean population value of 50 and a standard deviation (SD) of 10.We classify cases into: severe disability (>2.5 SD below the mean,i.e. MCS < 25); moderate disability (>1.5–2.5 SD below the mean,i.e. MCS = 25–34.9); and mild disability (> 0.5–1.5 SD below the mean,i.e. MCS = 35–44.9). The proportion of cases in each severity categoryis multiplied by the appropriate DW for the category to get an averageDW for those in each group. The resulting baseline DWs for GAD are:consulted = 0.22; consulted and received non-EBM = 0.20; and did notconsult = 0.19. For panic disorder, these are: consulted = 0.28; consultedand received non-EBM = 0.21; and did not consult = 0.12. Note that thehigher baseline disability weight in the population consulting for anxietydisorders is merely a result of the fact that more severely affectedindividuals are more likely to seek care.

Determining the reduction in YLD with treatment

The reduction in disability severity is modelled using the effect sizeand both the ‘conversion factor method’ and the ‘survey severitymethod’ to translate the effect size into a reduction in the DW. For the‘conversion factor method’ we multiply the effect size by the DW con-version factor for panic disorder/agoraphobia. This conversion factor isan average change in the DALY disability weights for the equivalent ofa standard deviation change in severity for the particular mental dis-order [25]. For the ‘survey severity method’ the effect size is applieddirectly to the Mental Component Score, which was used to determinethe average DW at baseline. The severity of respondents is then reclas-sified and a new average DW calculated. The difference in average DWis the change attributed to treatment [17].

The effect size for PD interventions is derived from a meta-analysisthat reports effect sizes for CBT, SSRIs and TCAs [16]. The effect sizefor TCAs (0.61) differed markedly from a previous meta-analysis(0.47) [26], so we decided to pool results from RCTs in both meta-analyses to obtain an overall effect size (Table 1). For cases notadherent with treatment, no reduction in DW has been modelled(although they do incur costs of the treatment provided).

Meta-analysis for GAD

For the meta-analysis of RCTs for GAD, the inclusion criteria weresubjects aged 18 years and older, a DSM-III-R or DSM-IV diagnosis ofGAD [27], and reporting of continuous measures (with both means andstandard deviations). Five small RCTs (total intervention n = 84 partic-ipants) met the inclusion criteria for CBT [28–32] and two larger RCTs(total intervention n = 435) met the inclusion criteria for venlafaxine[18,19].

L. HEUZENROEDER, M. DONNELLY, M.M. HABY, C. MIHALOPOULOS, R. ROSSELL, R. CARTER, G. ANDREWS, T. VOS 605

Tabl

e 1.

Cri

tica

l par

amet

er v

alue

s, u

ncer

tain

ty d

istr

ibut

ions

and

sou

rces

of i

nfor

mat

ion

for

dete

rmin

ing

heal

th b

enefi

ts a

nd c

osts

Par

amet

erV

alu

esU

nce

rtai

nty

d

istr

ibu

tio

n

†

Info

rmat

ion

so

urc

e an

d a

ssu

mp

tio

ns

Pre

vale

nce

an

d c

urr

ent

pra

ctic

e

‡

Pre

vale

nce

of G

AD

1-ye

ar: 3

.03%

, SE

0.2

%

1-m

onth

: 1.9

9%, S

E 0

.2%

1-y

ear

prev

alen

ce o

f G

AD

and

sym

ptom

atic

in th

e pa

st 2

wee

ks: 1

.92%

, S

E 0

.2%

Nor

mal

NS

MH

WB

[4]

Pro

port

ion

who

con

sulte

d an

d re

ceiv

ed n

on-

EB

M u

nder

cur

rent

pra

ctic

e (G

AD

)28

.2%

, SE

2.8

%N

orm

alN

SM

HW

B [4

]

Pre

vale

nce

of P

D1-

year

: 1.3

4%, S

E 0

.12%

1-

mon

th: 0

.65%

, SE

0.0

1%N

orm

alN

SM

HW

B [4

]

Pro

port

ion

who

con

sulte

d an

d re

ceiv

ed n

on-

EB

M u

nder

cur

rent

pra

ctic

e (P

D)

18.6

%, S

E 0

.13%

Nor

mal

NS

MH

WB

[4]

Hea

lth

ben

efit

Effe

ct s

ize

(GA

D in

terv

entio

ns)

Ven

lafa

xine

: min

0.2

8, m

ean

0.44

, max

0.6

0 C

BT

: min

0.4

5, m

ean

0.77

, max

1.0

9Tr

iang

ular

RC

Ts

of V

enla

faxi

ne [1

8,19

] and

CB

T [2

8–32

]. M

inim

um a

nd m

axim

um v

alue

s ar

e fr

om 9

5% C

I.E

ffect

siz

e (P

D in

terv

entio

ns)

SS

RIs

: min

0.1

7, m

ean

0.32

, max

0.4

7 T

CA

s: m

in 0

.20,

mea

n 0.

53, m

ax 0

.85

CB

T: m

in. 0

.41,

mea

n 0.

66, m

ax 0

.91

Tria

ngul

arF

or C

BT

and

SS

RIs

[16]

: F

or T

CA

s: R

CT

s fr

om th

e m

eta-

anal

yses

of i

nter

vent

ions

for

pani

c di

sord

er [1

6,26

]D

isab

ility

wei

ght c

onve

rsio

n fa

ctor

(G

AD

& P

D)

min

0.1

04, m

ax 0

.171

Uni

form

San

ders

on

et a

l

. [25

]. T

he m

inim

um is

from

the

Tim

e Tr

ade

Off

met

hod

and

the

max

imum

is fr

om th

e V

isua

l Ana

logu

e S

cale

Red

uctio

n in

the

disa

bilit

y w

eigh

t for

GA

D u

sing

th

e su

rvey

sev

erity

met

hod

Ven

lafa

xine

: min

0.0

3, m

ean

0.06

, max

0.0

8 C

BT

: min

0.0

6, m

ean

0.10

, max

0.1

2Tr

iang

ular

Met

hods

det

aile

d in

Hab

y

et a

l

. [17

]

Red

uctio

n in

the

disa

bilit

y w

eigh

t for

PD

usi

ng

the

surv

ey s

ever

ity m

etho

dS

SR

Is: m

in 0

.05,

mea

n 0.

06, m

ax 0

.09

TC

As:

min

0.0

5, m

ean

0.09

, max

0.1

6 C

BT

: min

0.0

8, m

ean

0.11

, max

0.1

6

Tria

ngul

arM

etho

ds d

etai

led

in H

aby

et a

l

. [17

]

Adh

eren

ce w

ith G

AD

inte

rven

tions

:V

enla

faxi

ne: m

in 5

0%, m

ax 7

3%

CB

T: m

in 5

0%, m

ax 9

5%U

nifo

rmM

inim

um v

alue

is a

n es

timat

e. M

axim

um v

alue

s ar

e ba

sed

on

adhe

renc

e w

ith th

e in

terv

entio

n in

RC

Ts

of V

enla

faxi

ne [1

8,19

] an

d C

BT

[28–

32]

Adh

eren

ce w

ith P

D in

terv

entio

nsS

SR

Is: m

in 5

0%, m

ax 6

7%

TC

As:

min

50%

, max

69%

C

BT

: min

50%

, max

85%

Uni

form

Min

imum

val

ue is

an

estim

ate.

Max

imum

val

ues

are

base

d on

ad

here

nce

with

the

inte

rven

tion

in R

CT

s in

the

met

a-an

alys

es fo

r pa

nic

diso

rder

[16,

26]

Co

nsu

ltat

ion

s

Num

ber

of d

aily

pat

ient

con

tact

sB

y pu

blic

psy

chol

ogis

t: m

in 5

, mea

n 6,

max

7Tr

iang

ular

Est

imat

e ba

sed

on c

onsu

ltatio

nB

y pu

blic

psy

chia

tris

t per

ses

sion

: min

2.5

, mea

n 3,

m

ax 3

.5

Co

sts

Var

iatio

n fa

ctor

aro

und

cost

of c

onsu

ltatio

ns w

ith

publ

ic p

sych

olog

ist a

nd p

ublic

psy

chia

tris

t (c

ost t

o go

vern

men

t) a

nd p

rivat

e ps

ycho

logi

st

(cos

t to

patie

nt)

mea

n 1,

SE

0.1

Nor

mal

Uni

t cos

ts a

nd s

ourc

es o

f dat

a ar

e sh

own

in T

able

2. T

hese

cos

ts

are

assu

med

to v

ary

toge

ther

, i.e

. in

the

sam

e di

rect

ion.

On-

cost

s fo

r pu

blic

psy

chol

ogis

t or

publ

ic

psyc

hiat

rist (

%)

min

25,

mea

n 30

, max

35

Tria

ngul

arO

n-co

sts

incl

ude

dire

ct s

alar

y co

sts

and

corp

orat

e co

sts

Cos

t of n

on-E

BM

var

iatio

n fa

ctor

min

0.5

, mea

n 1,

max

1.5

Tria

ngul

arE

stim

ate

†

In a

uni

form

dis

trib

utio

n ev

ery

valu

e in

the

spec

ified

ran

ge h

as a

n eq

ual p

roba

bilit

y of

bei

ng c

hose

n in

eac

h ite

ratio

n of

the

sim

ulat

ion.

In a

tria

ngul

ar d

istr

ibut

ion,

the

grea

test

pro

babi

lity

of

bein

g ch

osen

is th

e va

lue

repr

esen

ting

the

top

of th

e tr

iang

le (

i.e. t

he m

ean

valu

e sh

own

in c

olum

n 2)

, whi

le th

e pr

obab

ility

of o

ther

val

ues

bein

g ch

osen

tape

rs o

ff to

war

ds th

e ex

trem

es o

f the

ba

se o

f the

tria

ngle

(i.e

. the

min

imum

and

max

imum

val

ues)

;

‡

assu

mpt

ions

mad

e fo

r pr

eval

ence

and

cur

rent

pra

ctic

e es

timat

es w

ould

affe

ct th

e to

tal n

umbe

r el

igib

le fo

r th

e in

terv

entio

n an

d he

nce

tota

l cos

ts a

nd b

enefi

t but

hav

e no

effe

ct o

n th

e co

st-e

ffect

iven

ess

ratio

.

606 INTERVENTIONS FOR ANXIETY DISORDERS

We first calculate an effect size for each study by averaging acrossrelevant continuous outcome measures related to anxiety and depres-sion and health-related quality of life. Within each study, the effect size(standardized mean difference) is calculated using Hedges’ g andpooled across studies using the random effects method of DerSimonianand Laird [33]. The intervention effect sizes (for GAD and PD) arepresented in Table 1.

Adherence

It was assumed that the completion rate of the treatment group in theRCTs for GAD [18,19,28–32] and panic [16,26] reflects the best pos-sible adherence with treatment. No longitudinal studies measuringadherence to the interventions were available for GAD or PD, so aminimum rate of 50% was used in the uncertainty analysis (Table 1).This was done to better reflect what could be expected under routinehealth service conditions where results may vary due to the effect ofcomorbidity, the motivation of clinicians and patients, the availabilityof skilled clinicians and the capacity to vary the intervention to suit theneeds of the patient.

Stage two: the second stage filter criteria

The filters chosen for assessment in the ACE study were: ‘strengthof evidence’, ‘equity’ ‘feasibility’ and ‘acceptability to stakeholders’.The filters are described in Table 2.

Assessment of costs

Unit costs and data sources are shown in Table 3. Costs thatwould have been incurred under current practice are subtracted fromthe intervention (and nonadherence) costs to obtain the incrementalcost.

Health-care seeking behaviour for those not adherent with treat-ment is unknown; as a consequence, the cost of nonadherence mustbe estimated. It is assumed that the cost of nonadherence is (onaverage) the same as the cost of non-EBM. The cost of nonadher-ence is important to the ICER because those who do not adhere to

treatment would be expected to incur some cost but receive nohealth benefit.

Uncertainty analysis

Simulation modelling techniques are used to allow the presentationof an uncertainty range around the health benefits, costs and ICERS(Table 1) [34]. @RISK software [35] was used to conduct Monte Carlosimulations, which allow multiple recalculations of a spreadsheet, eachtime choosing a value from the specified distribution for each inputvariable (shown in Table 1). We use 2000 iterations for each of the twomethods for translating effect size into a change in the DW (i.e. the con-version factor method and the survey severity method). Thus, the finalresults (Tables 4,5) are based on the 2000 + 2000 iterations. Medianvalues were calculated because results are not normally distributed. Theranges presented can be interpreted as the range within which the trueresult lies with 95% certainty.

Results

Cognitive behavioural therapy provided by a public psychologist isthe most cost-effective intervention for treating both GAD and PD.Treatment of GAD with venlafaxine provides a similar ICER as CBTby other providers (Table 4). Tricyclic antidepressants are the secondmost cost-effective option for panic disorder, followed by CBT by otherproviders and SSRIs (Table 5). Within a 95% uncertainty range, all ofthe intervention options have ICERs not greater than A$40 000 forGAD and A$55 000 for PD.

For both GAD and PD, the major contributors to uncertainty aroundthe ICERs for all interventions (CBT, SNRI, SSRI and TCAs) are:the effect size; the reduction in disability weight; and (for CBT only)the variation factor around the cost of consulting private psychologists,private psychiatrists and general practitioners.

The results of the second-filter criteria assessment are presented inTable 6. A main outcome was the elucidation of issues surrounding theavailability and distribution of an adequate workforce for CBT inter-ventions. The drug interventions are likely to be more feasible,although possibly less acceptable, due to concerns about side-effects.

Table 2. The second stage filter criteria

Strength of evidence:

We used the NHMRC ‘Designation of levels of evidence’ but also highlight potential threats to the quality of the evidence by identifying possible biases and issues that affect the generalisability of the evidence. Based on the level and quality of evidence we categorized the strength of the evidence of effectiveness (or ineffectiveness) as: ‘sufficient’, ‘limited’ or ‘inconclusive’ [17]. This criterion is also addressed in the uncertainty modelling. For example, if the evidence on effectiveness is weak, large uncertainty is put around the size of the impact measure.

Equity:

This criterion addressed the capacity of the intervention to affect inequity in the distribution of the mental disorder, as well as highlighting issues that may affect access to, or utilization of, the intervention. The special needs groups considered include those with a lower socio-economic status, non-English speaking background, Aboriginal and Torres Strait Islanders or rural/remote residence.

Feasibility:

This criterion addressed issues such as the availability of an appropriate workforce with appropriate training to conduct the intervention, whether the intervention could be implemented under current institutional arrangements, and ease of implementation.

Acceptability to stakeholders:

This criterion referred to the anticipated acceptability of the proposed interventions to the various stakeholders affected by the intervention. Stakeholders include consumers and their family/carers, clinicians, policy makers, the general community and third-party funders.

L. HEUZENROEDER, M. DONNELLY, M.M. HABY, C. MIHALOPOULOS, R. ROSSELL, R. CARTER, G. ANDREWS, T. VOS 607

Discussion

Economic analysis raises important issues as to whatconstitutes ‘value-for-money’. It is not uncommon for athreshold ICER (or ‘shadow price’) to be set as a guideto assist decision-making. In ACE–MH, for example, anICER of A$50 000 per DALY has been used. However,this should not be over-interpreted or taken out of con-text. It is important to reflect, for example, on how wellthe ICER captures the various dimensions of ‘benefit’ inmental health. The second stage filters are designed toallow the ICERs to be placed within a broader decisioncontext.

Cognitive behavioural therapy is a cost-effectiveintervention for both GAD and PD, particularly whenprovided by publicly funded psychologists. It is likely(

≥

73% chance) that the ICERs for all interventions will

be below our threshold of A$50 000 per DALY saved.The SNRI for GAD has a similar ICER as CBT (byproviders other than public psychologists). However, ithas lower efficacy than CBT, resulting in lower totalhealth benefit.

For both disorders, the assumption was made thatdrugs would be obtained on the PBS. However, GADis not an indication for obtaining venlafaxine on thePBS (although depression, highly comorbid withGAD, is). Additionally, it is possible that some PDpatients will purchase other antidepressants not listedon the PBS (such as the SSRIs sertraline or fluvoxam-ine, or the TCA clomipramine). Had the model costeddrug interventions assuming no PBS benefit, overallcost-effectiveness would remain unchanged, althoughthe government would not contribute to payment fordrugs.

Table 3. Summary of unit cost information, data sources and assumptions

Element costed Cost to govt (A$)

Cost to patient (A$)

Source Assumptions

1 month supply of venlafaxine

†

$48.67 $9.99 PBS data from DHA Patients take their medicine as prescribed.

1 month supply of paroxetine

$67.55 $19.24

1 month supply of imipramine

$32.60 $19.00

1 GP visit of 20–40 mins

$39.51 $1.87 MBS data from DHA MBS item 36 – for initial visit and/or for referral/diagnosis

1 GP visit of < 20 mins $21.88 $2.21 MBS data from DHA MBS item 23 – for second and subsequent visits for SNRI intervention and for non-EBM

1 psychiatrist visit of 45–75 min

$117.02 $16.47 MBS data from DHA MBS item 306 – for CBT, non-EBM and for initial visit for SNRI, SSRI and TCA interventions

1 psychiatrist visit of 15–30 min

$56.38 $5.95 MBS data from DHA MBS item 302 – subsequent visits for SNRI, SSRI and TCA interventions

1 session of 46–60 mins with a private psychologist

$0 $115.00 APS APS recommended fee is A$161 (as of July 2001) but personal communication with the Executive Director suggests the fee most commonly charged is A$110–120.

1 session of 60 mins with a public pscyhologist

$47.05 $0 Base salary from Victorian Hospitals’ Industrial Association

Grade 3, Year 2 psychologist. Salary effective from 1 July 2000: A$1085.80 per week. On-costs of 30% have been added. The psychologist has 6 patient contacts per day with the remainder of the working day used for preparation, administration, professional development, etc.

1 session of 60 mins with a public psychiatrist

$129.64 $0 Base salary from the Austin & Repatriation Medical Centre Human Resources

The psychiatrist is paid as a visiting medical officer. An average salary per session has been used from the range: Specialist to Senior Specialist. Salary effective from 1 July 2000: A$260.76 to A$337.60 per session. The psychiatrist has 3 patient contacts per session (of 3.5 h). On-costs of 30% have been added.

†

Based on the wieghted average of the cost for doses 75 mg and 150 mg per day, as determined by the proportion of preseciptions made for each dose. APS, Australian Psychological Society; DHA, Australian Department of Health and Ageing; MBS, Medicare Benefits Schedule; PBS, Pharmaceutical Benefits Scheme; SNRI, serotonin and noradrenaline reuptake inhibitor; EBM, evidence-based medicine; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant.

608 INTERVENTIONS FOR ANXIETY DISORDERS

Tabl

e 4.

The

incr

emen

tal b

enefi

ts, c

osts

and

incr

emen

tal c

ost-

effe

ctiv

enes

s ra

tio

of c

ogni

tive

beh

avio

ur th

erap

y an

d th

e se

roto

nin

and

nora

dren

alin

e re

upta

ke in

hibi

tors

for

the

man

agem

ent o

f gen

eral

ized

anx

iety

dis

orde

r, co

mpa

red

wit

h cu

rren

t pra

ctic

e

Pri

vate

psy

cho

log

ist

Pu

blic

psy

cho

log

ist

Pri

vate

psy

chia

tris

tP

ub

lic p

sych

iatr

ist

SN

RI

Hea

lth

ben

efit

YLL

00

00

0Y

LD72

00 (

4300

–12

000)

7200

(43

00–1

2 00

0)72

00 (

4300

–12

000)

7200

(43

00–1

2 00

0)33

00 (

1900

–510

0)D

ALY

s72

00 (

4300

–12

000)

7200

(43

00–1

2 00

0)72

00 (

4300

–12

000)

7200

(43

00–1

2 00

0)33

00 (

1900

–510

0)

Co

st (

A$

Mill

ion

s)

†

Gov

ernm

ent

– 6.

5 (–

11 to

–2.

2)59

(42

–79)

160

(120

–190

)17

0 (1

20–2

30)

70 (

56–8

7)P

atie

nt15

0 (1

10–2

00)

– 8.

9 (–

14 to

–5.

1)14

(9.

2–19

)–

8.9

(–14

to –

5.1)

6.5

(3.6

–10)

Tota

l14

0 (1

00–1

90)

50 (

34–6

9)17

0 (1

30–2

10)

160

(120

–220

)77

(60

–96)

ICE

R (

A$

Th

ou

san

d/D

ALY

)

20 (

12–3

3)6.

9 (4

.0–1

2)23

(15

–38)

23 (

14–3

8)23

(16

–40)

Pro

port

ion

of it

erat

ions

bel

ow A

$50,

000/

DA

LY s

aved

‡

100

100

100

100

100

Non

-incr

emen

tal c

ost-

effe

ctiv

enes

s ra

tio

(A$

Tho

usan

ds/D

ALY

)

§

28 (

17–5

6)12

(7–

25)

32 (

20–6

3)31

(19

–63)

30 (

20–5

1)

Val

ues

are

med

ians

. Fig

ures

in b

rack

ets

show

the

95%

unc

erta

inty

inte

rval

; YLL

, yea

rs o

f life

lost

; YLD

, yea

rs li

ved

with

dis

abili

ty; D

ALY

s, d

isab

ility

adj

uste

d lif

e ye

ars

(YLL

+ Y

LD);

ICE

R, i

ncre

men

tal c

ost-

effe

ctiv

enes

s ra

tio; C

BT,

cog

nitiv

e be

havi

our

ther

apy;

SN

RI,

sero

toni

n an

d no

radr

enal

in r

eupt

ake

inhi

bito

r; G

AD

, gen

eral

ized

anx

iety

di

sord

er;

†

this

is th

e co

st o

f the

inte

rven

tion

redu

ced

by th

e co

st o

f non

-EB

M n

ot g

iven

;

‡

calc

ulat

ed fr

om th

e 40

00 it

erat

ions

gen

erat

ed in

the

unce

rtai

nty

anal

ysis

for

the

cost

-ef

fect

iven

ess

ratio

;

§

calc

ulat

ed fr

om th

e D

ALY

s sa

ved

and

inte

rven

tion

cost

s, i.

e. w

ithou

t the

cos

t sav

ings

from

non

-EB

M n

ot g

iven

; neg

ativ

e va

lues

are

sav

ings

from

the

redu

ctio

n in

non

-EB

M tr

eatm

ents

. Cos

t sav

ings

can

res

ult f

rom

a r

educ

tion

in to

tal c

ost o

f hea

lth s

ervi

ces

and/

or fr

om a

shi

fting

of c

osts

bet

wee

n pr

ovid

ers,

i.e.

gov

ernm

ent

vers

us

pat

ient

.

Tabl

e 5.

The

incr

emen

tal b

enefi

ts, c

osts

and

cos

t-ef

fect

iven

ess

of c

ogni

tive

beh

avio

ur th

erap

y an

d an

tide

pres

sant

s fo

r th

e m

anag

emen

t of p

anic

dis

orde

r, co

mpa

red

wit

h cu

rren

t pra

ctic

e

CB

TA

nti

dep

ress

ants

P

riva

te p

sych

olo

gis

tP

ub

lic p

sych

olo

gis

tP

riva

te p

sych

iatr

ist

Pu

blic

psy

chia

tris

tS

SR

IsT

CA

sH

ealt

h b

enefi

t

YLL

00

00

00

YLD

1200

(63

0–19

00)

1200

(63

0–19

00)

1200

(63

0–19

00)

1200

(63

0–19

00)

590

(250

–840

)87

0 (3

40–1

500)

DA

LYs

1200

(63

0–19

00)

1200

(63

0–19

00)

1200

(63

0–19

00)

1200

(63

0–19

00)

590

(250

–840

)87

0 (3

40–1

500)

Co

st (

A$

mill

ion

s)

†

Gov

ernm

ent

–1.9

(–

3.3

to .7

)12

(7.

5–17

)32

(24

–40)

36 (

23–5

1)21

(18

–24)

13 (

10–1

5)P

atie

nt28

.5 (

20–4

0)–

4.6

(– 7

.5 to

–2.

5)0.

2 (–

2.1

–2.1

)–

4.6

(– 7

.5 to

–2.

5)1.

5 (–

0.4

–27)

1.9

(0.9

–3.8

)To

tal

27 (

18–3

8)7.

0 (3

.2–1

2.1)

33 (

24–4

1)31

(20

–46)

23 (

18–2

7)15

(11

–19)

ICE

R (

A$

Th

ou

san

d/D

ALY

)

26 (

15–4

5)6.

8 (2

.9–1

4)27

(19

–48)

30 (

18–5

5)38

(27

–89)

17 (

9.7–

42)

Pro

port

ion

of it

erat

ions

bel

ow

A$5

0 00

0/D

ALY

sav

ed

‡

99%

100%

98%

95%

73%

99%

Non

-incr

emen

tal c

ost-

effe

ctiv

enes

s ra

tio (

A$

Tho

usan

ds/D

ALY

)

§

32 (

21–5

8)15

(10

–29)

36 (

23–6

6)36

(26

–65)

78 (

48–1

47)

30 (

18–7

3)

Val

ues

are

med

ians

; figu

res

in b

rack

ets

show

the

95%

unc

erta

inty

inte

rval

; YLL

, yea

rs o

f life

lost

; YLD

, yea

rs li

ved

with

dis

abili

ty; D

ALY

s, d

isab

ility

adj

uste

d lif

e ye

ars

(YLL

+ Y

LD);

ICE

R, i

ncre

men

tal c

ost-

effe

ctiv

enes

s ra

tio; S

SR

I, se

roto

nin

reup

take

inhi

bito

r; T

CA

, tric

yclic

ant

idep

ress

ant

†

.Thi

s is

the

cost

of t

he in

terv

entio

n re

duce

d by

the

cost

of n

on-E

BM

no

t giv

en;

‡

calc

ulat

ed fr

om th

e 40

00 it

erat

ions

gen

erat

ed in

the

unce

rtai

nty

anal

ysis

for

the

ICE

R;

§

calc

ulat

ed fr

om th

e D

ALY

s sa

ved

and

inte

rven

tion

cost

s, i.

e. w

ithou

t the

cos

t sa

ving

s fr

om n

on-E

BM

not

giv

en; n

egat

ive

valu

es a

re s

avin

gs fr

om th

e re

duct

ion

in n

on-E

BM

trea

tmen

ts. C

ost s

avin

gs c

an r

esul

t fro

m a

red

uctio

n in

tota

l cos

t of h

ealth

ser

vice

s an

d/or

from

a s

hifti

ng o

f cos

ts b

etw

een

prov

ider

s, i.

e. g

over

nmen

t

vers

us

pat

ient

.

L. HEUZENROEDER, M. DONNELLY, M.M. HABY, C. MIHALOPOULOS, R. ROSSELL, R. CARTER, G. ANDREWS, T. VOS 609

A lack of economic evaluations analyzing interven-tions for anxiety disorders limits comparison of theresults of this analysis. The only economic evaluationcomparing interventions for panic disorder was a cost-benefit analysis by Otto

et al

. [36] in the US, withtreatment benefits and costs based on a trial undertakenby the authors. Costs were calculated for the servicesand/or drugs required to maintain for one year a one-point increase in the Clinician Global Impression ofSeverity scale. Pharmacotherapy was considerably moreexpensive (US$1153) than individual CBT (US$646)and particularly group CBT (US$248). The authors con-cluded that, based on available evidence of cost, accept-ability and tolerability, and treatment outcome, CBTshould be the initial mode of treatment. However, as thiswas a non-randomised, non-controlled trial with nouncertainty analysis, the results are not directly compa-rable with the current study.

A recent cost-effectiveness analysis undertaken byIssakidis

et al

. [37] examines disability averted by pro-viding a mix of ‘optimal therapies’ to those receiving‘current therapies’ for a mix of anxiety disorders.Optimal therapies were actually treatment ‘packages’

which emphasized CBT as optimum treatment for mostpatients, but also included pharmacological treatment foraround one-third of patients. The cost-effectiveness ofstand-alone CBT or pharmacotherapy was not analyzed.Also in contrast to the current study, the comparator was‘no treatment’, so the placebo effect was added to theeffect size of these interventions, making the healthgains much greater. Therefore, the study results are notcomparable with this analysis of specific interventionsfor GAD/PD.

The strengths and limitations of the methods commonacross the ACE–MH project are presented in detail inHaby

et al

. [17]. As discussed in that introductory paper,the calculation of health gain is the most difficult meth-odological issue in the ACE–MH project. One limitingfactor specific to this analysis is the absence of longitu-dinal community surveys of the Australian population,which would provide a greater understanding of thenature and duration of GAD and PD and associated long-term treatment patterns and health service utilization.The available treatment research fails to measure theeffect of comorbidity. People with comorbidities are oftenexcluded and/or the papers do not present the results

Table 6. Results from assessment of second filters

Filter CBT intervention Antidepressants

Evidence Sufficient evidence of adequate quality, noting however, that there are:

●

few trials in NESB groups

●

few trials for providers other than psychologists

●

no trials amongst the indigenous population

●

only one trial in children <13 years

Sufficient evidence of adequate quality.

Equity Moderate equity concerns require attention, i.e.

●

appropriateness for minority groups (e.g. NESB, indigenous)

●

access for rural/remote consumers and in outer metropolitan areas (computer-based CBT may address this issue)

●

inequity in access if ‘user pays’ (e.g. for private providers)

No important equity issues for TCAs or SSRIs. However, access concerns for lower socio-economic groups when venlafaxine is prescribed solely for GAD and therefore not through the PBS.

Feasibility Possible but challenging to implement in short-term. Issues include:

●

ensuring an adequate workforce, i.e. appropriately trained and accredited providers; adequate distribution

●

ability of health funding to enable adequate access via primary care

●

development of roll-out arrangements (cost-effectiveness assumes ‘steady-state’ operation)

Feasible within current workforce and institutional arrangements.

Acceptability Concerns include:

●

cost to consumers if private providers

●

acceptance of treatment by clinicians and consumers

●

acceptance of a shift towards non-pharmacological treatments

Concerns include:

●

consumer concern about side-effects of drugs where treatment is long-term

●

reluctance by clinicians to prescribe TCAs due to perception of greater side-effects, and pressure to prescribe newer drugs (such as SSRIs)

CBT, cognitive behaviour therapy; TCA, tricyclic antidepressant; SSRI, selective serotonin reuptake inhibitor; GAD, generalized anxiety disorder; NESB, non-English speaking background.

610 INTERVENTIONS FOR ANXIETY DISORDERS

stratified by subgroups according to comorbidity. There-fore, use of these results may not be generalizable to allthose with comorbidities. However, the results are likelyto be a good reflection of what would happen ‘on aver-age’ and allow meaningful comparisons between treat-ments.

Particular to GAD, the lack of studies examining theeffectiveness of interventions over the long-term makesit impossible to know whether pharmaceutical or psy-chological interventions are the most appropriate way tomanage the chronic nature of this disorder. A recentreview of studies exploring long-term pharmacologicaltreatment of GAD concluded that evidence is extremelylimited; the few studies identified were methodologi-cally weak, with no placebo control and based on pre-DSM-III-R-defined GAD [38].

By contrast, there is no evidence for long-term gainsfollowing drug treatment for PD. In the Gould meta-analysis, which mostly analyzed follow-up trials forimipramine, it was reported that the effect of pharmaco-logical treatments wore off once administration ceased.Moreover, the few available long-term and follow-upstudies for SSRIs report an overall worsening of symp-toms following the tapering off of medication [39–41].

The GAD and PD literature fails to explore the effi-cacy of combined pharmaceutical and psychologicalinterventions [42,43]. Factorial design studies arerequired to estimate the effect of CBT relative topharmacological interventions alone and in combinationwith psychological interventions. In addition, there is aclear need for long-term studies comparing SNRIsagainst similar antidepressants such as the SSRIs (e.g.paroxetine), the older and cheaper TCAs (e.g. imi-pramine) and buspirone. However, funding would haveto come from public sources, as pharmaceutical com-panies would not be expected to benefit commerciallyfrom performing such research.

While CBT by a public psychologist was the mostcost-effective intervention for both anxiety disorders, thesecond filter analysis revealed issues that may impact onits implementation. Greater use of publicly funded psy-chologists will require attention to ensuring an adequateworkforce, particularly in outer metropolitan and ruralregions. It is important to note here that our cost-effectiveness analyses assume steady-state operation soimplementation costs are not included in the analyses.

Another possibility for the CBT intervention is to funda mix of providers including other suitably trained healthprofessionals (social workers, nurses, GPs) as currentlybeing piloted in the ‘Better outcomes in mental healthcare’ initiative [44]. However, use of providers that arenot adequately trained may decrease the effectiveness,and therefore cost-effectiveness, of the intervention.

Thus, attention to training and accreditation will berequired if similar effectiveness is to be achieved.

A change in the cost of accessing different providerswill also change the cost-effectiveness of the interven-tion. We have modelled the same effectiveness for thedifferent providers but different costs, ranging from $47per session for a psychologist on a public salary to $133for a private psychiatrist (which includes the Medicarerebate and out-of-pocket cost to the patient). A mix ofproviders would result in a cost-effectiveness ratio some-where between that for a public psychologist and that fora private psychiatrist.

Access for outer metropolitan and rural regions is animportant consideration for most providers but the useof computer-based CBT may overcome this problem iffound to be as effective for anxiety disorders as fordepression [45]. In comparison to CBT, the healthsystem provides no barriers to accessing drug treatmentfor GAD or PD, although the fact that drug therapy ismore accessible (by both cost and availability of pro-viders) than psychological therapies, could be unaccept-able to consumers. However, an exception is venlafaxinefor GAD, which is not currently available on the PBS,which presents an equity issue for access to this inter-vention.

Caveat

The Assessing Cost-Effectiveness – Mental Health(ACE–MH) project was jointly funded by the AustralianDepartment of Health and Ageing, Mental Health andSuicide Prevention Branch and the Department ofHuman Services, Mental Health Branch, Victoria inrecognition of the importance of research into the cost-effectiveness of interventions in mental health treatmentand care. This work draws upon, but is also limited bythe available research and the assumptions necessary tocomplete the work.

The results of the analyses provide valuable material,likely to contribute to future policy deliberations by allservice providers. Conclusions drawn from the eco-nomic evaluations should be considered within thecontext of the second stage filter process, which qualifiesthe results taking into account issues of equity, feasi-bility, strength of evidence, and acceptability to stake-holders. This second stage filter process addresses someof the practical considerations required for changes inactual service practice.

Acknowledgements

Principal investigators for the project are: Theo Vos,Rob Carter and Gavin Andrews. Analyses draw on the

L. HEUZENROEDER, M. DONNELLY, M.M. HABY, C. MIHALOPOULOS, R. ROSSELL, R. CARTER, G. ANDREWS, T. VOS 611

National Survey of Mental Health and Wellbeing, forwhich unit record data was obtained from the AustralianBureau of Statistics (ABS) [4] and a revised scoringalgorithm to determine ICD-10 and DSM-IV diagnoseswas provided by Gavin Andrews. Information containedin this report on current prescribing behaviour of GPshas been drawn from data collected by the GeneralPractice Statistics and Classification Unit, the Universityof Sydney in collaboration with the Australian Instituteof Health and Welfare [6]. The average cost of varioustypes of medical attendances and the various forms ofSSRI were obtained from Medicare Benefits Scheduleand Pharmaceutical Benefits Scheme data from theDepartment of Health and Ageing. We thank GavinAndrews, Kristy Sanderson, Caroline Hunt and CathIssakidis for advice on various aspects of the analysis.

We thank members of the ACE–MH steering committeefor their input into the project: David Barton, GrahamBurrows (Chair), Sue Caleo, Vaughan Carr, DermotCasey, Joy Easton, William Hart, Helen Herrman, BarbaraHocking, Assen Jablensky, Anthony Jorm, Lyn Littlefield,Patrick McGorry, John McGrath, Paul Morgan, LornaPayne, Deb Podbury, Kristy Sanderson, Suzy Saw, BruceSingh, Bruce Tonge, Ruth Vine, Harvey Whiteford.

References

1. Mathers CD, Vos T, Stevenson C.

The burden of disease and injury in Australia

. Canberra: Australian Institute of Health and Welfare, 1999.

2. Vanin JR, Vanin SK. Blocking the cycle of panic disorder.

Postgraduate Medicine

1999; 105:141–146.3. World Health Organisation.

The ICD-10 classification of mental and behavioural disorders: clinical descriptions and diagnostic guidelines.

Geneva: Switzerland, 1992.4. Australian Bureau of Statistics.

Mental health and wellbeing: profile of adults, Australia 1997.

Canberra: Australian Bureau of Statistics, 1998.

5. Therapeutic Guidelines Limited.

Therapeutic guidelines: psychotropic.

4th edn. Melbourne: Therapeutic Guidelines Limited, 2000.

6. Britt H, Miller G, Knox S

et al.

General practice activity in Australia 2000–01. General Practice Series No. 8

. Canberra: Australian Institute of Health and Welfare, 2001.

7. Kapczinski F, Schmitt R, Lima M. The use of antidepressants for generalized anxiety disorder. In:

The Cochrane Library,: Issue 2

. Oxford: Update Software.

8. Sheehan DV. Attaining remission in generalized anxiety disorder: venlafaxine extended release comparative data.

Journal of Clinical Psychiatry

2001; 62(Suppl. 19):26–31.9. Gelenberg AJ, Lydiard RB, Rudolph RL, Aguiar L, Haskins JT,

Salinas E. Efficacy of venlafaxine extended-release capsules in nondepressed outpatients with generalized anxiety disorder: a 6-month randomized controlled trial.

Journal of the American Medical Association

2000; 283:3082–3088.10. Rickels K, Pollack MH, Sheehan DV, Haskins JT. Efficacy of

extended-release venlafaxine in nondepressed outpatients with generalized anxiety disorder.

American Journal of Psychiatry

2000; 157:968–974.

11. Silverstone PH, Salinas E. Efficacy of venlafaxine extended release in patients with major depressive disorder and comorbid generalized anxiety disorder.

Journal of Clinical Psychiatry

2001; 62:523–529.

12. Pollack MH, Zaninelli R, Goddard A

et al.

Paroxetine in the treatment of generalized anxiety disorder: results of a placebo-controlled, flexible-dosage trial.

Journal of Clinical Psychiatry

2001; 62:350–357.13. Rocca P, Fonzo V, Scotta M, Zanalda E, Ravizza L. Paroxetine

efficacy in the treatment of generalized anxiety disorder.

Acta Psychiatrica Scandinavica

1997; 95:444–450.14. Rudolph L, Entsuah R, Chitra R. A meta-analysis of the effects

of venlafaxine on anxiety associated with depression.

Journal of Clinical Psychopharmacology

1998; 18:136–144.15. Caswell A ed.

MIMS

Sydney: MediMedia Australia, 2001.16. Royal Australian and New Zealand College of Psychiatrists

Clinical Practice Guideline Team for Panic Disorder and Agoraphobia. Australian and New Zealand clinical practice guidelines for the treatment of panic disorder and agoraphobia.

Australian and New Zealand Journal of Psychiatry

2003; 37:641–656.

17. Haby M, Carter R, Mihalopoulos C, Magnus A, Andrews G, Vos T. Assessing Cost-Effectiveness – Mental Health: introduction to the study and methods.

Australian and New Zealand Journal of Psychiatry

2004; 38:569–578.18. Allgulander C, Hackett D, Salinas E. Venlafaxine extended

release ER in the treatment of generalised anxiety disorder: twenty-four-week placebo-controlled dose-ranging study.

British Journal of Psychiatry

2001; 179:15–22.19. Davidson JR, DuPont RL, Hedges D, Haskins JT. Efficacy,

safety, and tolerability of venlafaxine extended release and buspirone in outpatients with generalized anxiety disorder.

Journal of Clinical Psychiatry

1999; 60:528–535.20. Australian Bureau of Statistics.

Census data 2000: population by age and sex, Australian states and territories.

Time series spreadsheet, 32010.0: Table 9. Canberra: Australian Bureau of Statistics, 2001.

21. Angst J, Vollrath M. The natural history of anxiety disorders.

Acta Psychiatrica Scandinavica

1991; 84:446–452.22. Keller M. The long-term clinical course of generalised anxiety

disorder.

Journal of Clinical Psychiatry

2002; 63(Suppl. 8): 11–16.

23. Yonkers K, Dyck I, Warwhaw M, Keller M. Factors predicting the clinical course of generalised anxiety disorder.

British Journal of Psychiatry

2000; 176:544–549.24. Stouthard M, Essink-Bot M, Bonsel G

et al.

Disability weights for diseases in the Netherlands

. Rotterdam: Department of Public Health, Erasmus University, 1997.

25. Sanderson K, Andrews G, Corry J, Lapsley H. Using the effect size to model change in preference values from descriptive health states.

Quality of Life Research

2004; 13:in press.

26. Gould RA, Otto MW, Pollack MH. A meta-analysis of treatment outcome for panic disorder.

Clinical Psychological Review

1995; 15:819–844.

27. American Psychiatric Association.

Diagnostic and statistical manual of mental disorders (DSM-IV)

4th edn. Washington, DC: American Psychiatric Press, 1994.

28. Borkovec TD, Costello E. Efficacy of applied relaxation and cognitive-behavioral therapy in the treatment of generalized anxiety disorder.

Journal of Consulting and Clinical Psychology

1993; 61:611–619.

29. Butler G, Fennell M, Robson P, Gelder M. Comparison of behavior therapy and cognitive behavior therapy in the treatment of generalized anxiety disorder.

Journal of Consulting and Clinical Psychology

1991; 59:167–175.

612 INTERVENTIONS FOR ANXIETY DISORDERS

30. Durham RC, Murphy T, Allan T, Richard K, Treliving LR, Fenton GW. Cognitive therapy, analytic psychotherapy and anxiety management training for generalised anxiety disorder.

British Journal of Psychiatry

1994; 165:315–323.31. Ladouceur R, Dugas M, Freeston M, Gagnon F, Thibodeau N.

Efficacy of a cognitive behavioural treatment for generalized anxiety disorder: evaluation in a controlled clinical trial.

Journal of Consulting Clinical Psychology

2000; 68:957–964.32. White J, Keenan M. Stress control: a controlled comparative

investigation of large group therapy for generalized anxiety disorder.

Behavioural Psychotherapy

1992; 20:97–114.33. Deeks J, Altman D, Bradburn M. Statistical methods for

examining heterogeneity and combining results from several studies in meta-analysis. In: Egger M, Smith G, Altman D eds.

Systematic reviews in health care: meta-analysis in context

, 2nd edn. London: BMJ Publishing Group, 2001; 285–312.

34. Drummond M, McGuire A eds.

Economic evaluation in health care: merging theory with practice

. New York: Oxford University Press, 2001.

35. Palisade Corporation.

@RISK [computer program] Risk Analysis Add-in for Microsoft Excel

, Version 4. Newfield, NY: Palisade Corporation, 2001.

36. Otto MW, Pollack M, Maki KM. Empirically supported treatments for panic disorder: costs, benefits and stepped care.

Journal of Consulting and Clinical Psychology

2000; 68:556–563.

37. Issakidis C, Sanderson K, Corry J, Andrews G, Lapsley H. Modelling the population cost-effectiveness of current and evidence based optimal treatment for anxiety disorders.

Psychological Medicine

2004; 34:19–35.

38. Mahe V, Balogh A. Long-term pharmacological treatment of generalized anxiety disorder.

International Clinical Psychopharmacology

2000; 15:99–105.39. Michelson D, Pollack M, Lydiard RB, Tamura R, Tepner R,

Tollefson G. Continuing treatment of panic disorder after acute response: randomised, placebo-controlled trial with fluoxetine. The Fluoxetine Panic Disorder Study Group.

British Journal of Psychiatry 1999; 174:213–218.

40. Rapaport MH, Wolkow R, Rubin A, Hackett E, Pollack M, Ota KY. Sertraline treatment of panic disorder: results of a long-term study. Acta Psychiatrica Scandinavica 2001; 104:289–298.

41. de Beurs E, van Balkom AJ, Van Dyck R, Lange A. Long-term outcome of pharmacological and psychological treatment for panic disorder with agoraphobia: a 2-year naturalistic follow-up. Acta Psychiatrica Scandinavica 1999; 99:59–67.

42. Lader MH, Bond AJ. Interaction of pharmacological and psychological treatments of anxiety. British Journal of Psychiatry 1998; (Suppl.):42–48.

43. Perry S. Combining antidepressants and psychotherapy: rationale and strategies. Journal of Clinical Psychiatry 1990; 51(Suppl.):16–20.

44. Mental Health and Suicide Prevention Branch. Better outcomes in mental health care initiative. Canberra: Australian Department of Health and Ageing, 2002. (Available at: http://www.health.gov.au/hsdd/mentalhe/boimhc/index.htm Last accessed 12 November 2003.)

45. Selmi P, Klein M, Greist J, Sorrell S, Erdman H. Computer-administered cognitive-behavioural therapy for depression. American Journal of Psychiatry 1990; 147:51–56.