Available via license: CC BY 4.0
Content may be subject to copyright.
Articles
www.thelancet.com/psychiatry Vol 3 May 2016
415
Scaling-up treatment of depression and anxiety: a global
return on investment analysis
Dan Chisholm, Kim Sweeny, Peter Sheehan, Bruce Rasmussen, Filip Smit, Pim Cuijpers, Shekhar Saxena
Summary
Background Depression and anxiety disorders are highly prevalent and disabling disorders, which result not only in
an enormous amount of human misery and lost health, but also lost economic output. Here we propose a global
investment case for a scaled-up response to the public health and economic burden of depression and anxiety
disorders.
Methods In this global return on investment analysis, we used the mental health module of the OneHealth tool to
calculate treatment costs and health outcomes in 36 countries between 2016 and 2030. We assumed a linear increase
in treatment coverage. We factored in a modest improvement of 5% in both the ability to work and productivity at
work as a result of treatment, subsequently mapped to the prevailing rates of labour participation and gross domestic
product (GDP) per worker in each country.
Findings The net present value of investment needed over the period 2016–30 to substantially scale up eff ective
treatment coverage for depression and anxiety disorders is estimated to be US$147 billion. The expected returns to
this investment are also substantial. In terms of health impact, scaled-up treatment leads to 43 million extra years of
healthy life over the scale-up period. Placing an economic value on these healthy life-years produces a net present
value of $310 billion. As well as these intrinsic benefi ts associated with improved health, scaled-up treatment of
common mental disorders also leads to large economic productivity gains (a net present value of $230 billion for
scaled-up depression treatment and $169 billion for anxiety disorders). Across country income groups, resulting
benefi t to cost ratios amount to 2·3–3·0 to 1 when economic benefi ts only are considered, and 3·3–5·7 to 1 when the
value of health returns is also included.
Interpretation Return on investment analysis of the kind reported here can contribute strongly to a balanced
investment case for enhanced action to address the large and growing burden of common mental disorders worldwide.
Funding Grand Challenges Canada.
Copyright © Chisholm et al. Open Access article distributed under the terms of CC BY.
Introduction
Worldwide, investments in mental health are very
meagre. Data from WHO’s Mental Health Atlas 2014
survey1 suggest that most low-income and middle-income
countries spend less than US$2 per year per person on
the treatment and prevention of mental disorders
compared with an average of more than $50 in high-
income countries. As a result of this limited investment
in public mental health, a substantial gap exists between
the need for treatment and its availability. This large
treatment gap aff ects not just the health and wellbeing of
people with mental disorders and their families, but also
has inevitable consequences for employers and
governments as a result of diminished productivity at
work, reduced rates of labour participation, foregone tax
receipts, and increased health and other welfare
expenditures. Findings of several national and
international studies2–5 have shown the enormous
economic challenge these disorders pose to communities
and society at large as a result of foregone production
and consumption opportunities as well as health and
social care expenditures. In 2010, worldwide, an
estimated US$2·5–8·5 trillion in lost output was
attributed to mental, neurological and substance use
disorders, depending on the method of assessment
used.2 This sum is expected to nearly double by 2030 if a
concerted response is not mounted.2 In view of this
concern, the promotion of mental health and wellbeing
have been explicitly included in the United Nations’
2015–30 Sustainable Development Goals.6
Cost-eff ectiveness studies have largely restricted
themselves to a consideration of the specifi c
implementation costs and health outcomes of an
intervention, and have typically not extended to a full
estimation of the wider socioeconomic value of
investment in mental health innovation and service
scale-up. As shown in the Lancet’s Commission on
Investing in Health, elucidation and enumeration of
these wider economic and social benefi ts provides a
more comprehensive assessment of the returns on
investment.7 In particular, increasing attention and
emphasis is being given to extending valuation to also
include the intrinsic value of improved health (a so-called
full income approach to national accounting).7
Lancet Psychiatry 2016;
3: 415–24
Published Online
April 12, 2016
http://dx.doi.org/10.1016/
S2215-0366(16)30024-4
See Comment page 390
Department of Mental Health
and Substance Abuse, WHO,
Geneva, Switzerland
(D Chisholm PhD, S Saxena MD);
Victoria Institute of Strategic
Economic Studies, Melbourne,
VIC, Australia (K Sweeny PhD,
Prof P Sheehan PhD,
Prof B Rasmussen PhD); Trimbos
Institute (Netherlands
Institute of Mental Health and
Addiction), Utrecht,
Netherlands (Prof F Smit PhD);
and Department of Clinical,
Neuro and Developmental
Psychology, Vrije Universiteit
Amsterdam, Netherlands
(Prof F Smit, Prof P Cuijpers PhD)
Correspondence to:
Dr Dan Chisholm, Department of
Mental Health and Substance
Abuse, WHO, Geneva 1211,
Switzerland
ChisholmD@WHO.int
Articles
416
www.thelancet.com/psychiatry Vol 3 May 2016
Here we did a global return on investment analysis for
mental health in people aged 15 years and older focusing
on depression and anxiety disorders, which are the most
prevalent mental disorders. These disorders lead to large
losses in work participation and productivity, and yet
lend themselves to eff ective and accessible treatment as
part of an integrated programme of chronic disease
management.8–10
Methods
Analytical framework
Because depression and anxiety disorders represent a
public health challenge worldwide, we did a global
investment appraisal in low-income, middle-income, and
high-income countries. The 36 countries for which we
modelled costs and benefi ts of scaled-up treatment,
which span all six of WHO’s major regions, account for
80% of the world’s population and 80% of the global
burden of depression and anxiety disorders (appendix p 1).
Results for these countries were aggregated and reported
by income level (low, lower-middle, upper-middle, high).
We set the scale-up period at 2016–30, in line with the
timeline of the post-2015 Sustainable Development
Goals.
The economic and social benefi ts of good mental
health include both its intrinsic value (improved mental
health and wellbeing) and also its instrumental value, in
terms of being able to form and maintain relationships,
to work or pursue leisure interests, and to make decisions
in everyday life. To assess the value of these benefi ts, fi rst
we estimated the population in need in each country,
then established the health eff ects of scaled-up coverage
of eff ective intervention, and fi nally calculated the
economic eff ect of improved mental health outcomes in
terms of enhanced labour participation and productivity.
Panel 1 provides more detail on the health and economic
benefi ts captured in, and omitted from, the analysis. The
key outputs of the model are year-on-year estimates of
the total costs of treatment scale-up and system
strengthening (ie, the investment), increased healthy
life-years gained as a result of treatment (ie, health
return), the value associated with better health (ie, the
value of health returns), and enhanced levels of
productivity (ie, economic return). The stream of costs
incurred and benefi ts obtained between 2016 and 2030
were discounted at a rate of 3%, to give a net present
value. All costs and monetised benefi ts were expressed in
constant US$ for the year 2013.
Population and disease modelling
We used the mental health module of the inter-UN
agency OneHealth tool to estimate the number of people
with depression and anxiety disorders living in the
36 large countries until 2030. Estimates are based on UN
population projections and Global Burden of Disease
prevalence estimates for 2010.11,12 The global point
prevalence rate for anxiety disorders is 7·3%;13 for
depression it is 3·2% for men, and 5·5% for women.14
The OneHealth tool also links the epidemiology of
depression and anxiety disorders (prevalence, incidence,
remission, excess mortality, and disability weight)12–14 to
country-specifi c life tables, so that cases averted and
Research in context
Systematic review
We did a systematic review of studies published until Jan 1,
2015, of the eff ect of treatment of depression and anxiety
disorders on economic outcomes (return to work, absenteeism,
and presenteeism rates). We searched an existing database on
psychological treatments of depression, which has been
described in detail by Cuijpers and colleagues, and has been
used in a series of earlier published meta-analyses. We
identifi ed abstracts by combining terms indicative of
psychological treatment and depression (both medical subject
headings terms and text words; search terms listed in appendix
p 4). For this database, we examined 17 061 abstracts from
PubMed (4007 abstracts), PsycINFO (3147 abstracts), Embase
(5912 abstracts), and the Cochrane Central Register of
Controlled Trials (3995 abstracts). We included all randomised
trials comparing a psychological treatment with a control
condition (waiting list, care as usual, placebo), another
psychological treatment, pharmacotherapy, or combined
treatment. We excluded studies in adolescents, children, and
inpatients, and maintenance trials. We scrutinised all
440 studies identifi ed in this database for economic outcome
data. Although four studies had data for functioning at work,
they did not report suffi cient detail. Accordingly, we did an
additional search on May 21, 2105, to widen our search to
include anxiety disorders with greater emphasis on economic
outcomes in PubMed, EMBASE, PsycINFO and the Cochrane
Library (search terms listed in appendix p 5). We found few
useful data and these could not be synthesised
meta-analytically. The same conclusion was made in a similar
review of the scientifi c literature.
Interpretation
This analysis sets out a model linking the prevalence of
depression and anxiety disorders with expected health and
economic benefi ts of scaled-up treatment, including restored
labour participation and productivity. Results from the analysis
suggest that monetised benefi ts of better health and labour
force outcomes outweigh the costs of achieving them by
2·3–3·0 to 1 when economic benefi ts only are considered, and
3·3–5·7 to 1 when the value of health returns is also included.
Treatment of common mental disorders leads to improvements
in economic production and health outcomes. Clinicians should
increase the detection and management of people with
depression and anxiety disorders.
For the published
meta-analyses see http://www.
evidencebasedpsychotherapies.
org
For the OneHealth tool see
http://www.who.int/choice/
onehealthtool
See Online for appendix
Articles
www.thelancet.com/psychiatry Vol 3 May 2016
417
healthy life-years gained over time at the population level
can be estimated. Healthy life-years refl ect time spent by
the population in a particular state of health with a
known degree of disability. Estimation of healthy life-
years for depression took into account its association
with excess mortality (due to suicide and other causes of
death).14
Intervention eff ects, costs, and coverage
Intervention eff ects
We restricted the analysis of interventions within the
OneHealth tool to treatment because the evidence on
prevention of depression and anxiety is quite weak and of
uncertain generalisability to low-income and middle-
income country settings.15 In line with WHO’s Mental
Health Gap Action Programme (mhGAP) intervention
guide, modelled interventions included basic psycho-
social treatment for mild cases, and either basic or more
intensive psychosocial treatment plus anti depressant
drug for moderate to severe cases.16 Moderate to severe
cases of depression were split into fi rst-episode and
recurrent episode cases. We calculated the health eff ect
of treatment in terms of a proportionate improvement in
the rate of remission, equivalent to a shortening of the
duration of an episode of illness, and also, up to the point
of recovery, an improvement in the average level of
functioning as refl ected in the disability weight for the
disorder.8,10 The appendix shows the eff ect size estimates
and their derivation (appendix p 2); these take into
account partial response, the lag time between onset of
the disorder and treatment, and expected levels of non-
adherence in treated populations.
Intervention costs
We worked out total costs in a given year for a country
by multiplying resource use needs by their respective
unit costs to give a cost per case, which was then
multiplied by the total number of cases expected to
receive a particular intervention. Country-specifi c unit
costs of inpatient and outpatient care were taken from a
WHO database, adjusted to 2013 price levels.17 Treatment
costs relied on previous cost-eff ectiveness studies and
resource need profi les garnered from existing treatment
guidelines and costing studies.10,16,18,19 Key categories of
resource use were: medication: 6 months continual
antidepressant drug (generically produced fl uoxetine)
was included for moderate to severe cases; outpatient
and primary care: regular visits were needed for all
cases, ranging from four per case per year for basic
psychosocial treatment, up to 14–18 visits for moderate
to severe cases receiving antidepressant drug and
intensive psychosocial treatment (half of whom are
assumed to receive this on an individual basis, the other
half in groups); in line with the mhGAP intervention
guide, it is envisaged that this care and follow-up would
largely be undertaken in non-specialist health care
settings by doctors, nurses and psychosocial care
providers trained in the identifi cation, assessment, and
management of depression and anxiety disorders; and
inpatient care: few cases are expected to be admitted to
hospital (2–3% of moderate to severe cases only, for an
average length of stay of 14 days).
Additionally, we included an estimate of the expected
level of programme costs and shared health system
resources needed to deliver interventions as part of an
Panel 1: Health, economic, and social benefi ts of scaled-up treatment for depression
and anxiety disorders
Health eff ects
To establish the eff ect of treatment, we used rates of improved recovery or remission and
levels of functioning. Improved functioning translates into fewer life-years spent by the
population in a state of diminished health, whereas an increased rate of remission leads
to a decrease in the prevalence of these disorders over time. Depression is also associated
with an excess risk of premature mortality because of suicide and other causes of death.
We projected a reduction in excess mortality, amounting to an increase in healthy life
expectancy, as a result of averting cases of depression in the population. Although
depression and anxiety disorders are often comorbid with each other, and with a range of
other health disorders (eg, substance use disorder, other non-communicable diseases and,
in certain populations, in people with HIV/AIDS) we were not able to account for these
comorbidities in the analysis. Additionally, we were unable to capture the positive eff ect
of treatment on the mental and physical health of close family members, including
infants of mothers with perinatal depression, despite robust evidence that depression can
adversely aff ect infant attachment and subsequent child growth and cognitive
development.
Economic eff ects
A direct potential benefi t of successfully treating common mental disorders is a decrease in
overall health-care costs. Although interventions have their own costs, these can be more
than off set by a reduction in other services, notably hospital-based inpatient episodes or
outpatient visits. Reduced use of informal and indigenous health-care providers, such as
faith healers or traditional healers, is a further expected source of cost savings in many
countries. Estimation of the predicted extent of these cost off sets is very challenging at the
international level because it requires detailed information about both the varying level of
comorbidity across diverse populations and the typical use of non-intervention related
services. Accordingly, we did not explicitly consider such eff ects in our analysis. Similarly, we
did not have suffi cient information across countries to model the reduced need for other
welfare-related services potentially available to people with depression and anxiety
disorders, including unemployment benefi t or income support and social or disability
assistance. In the mainly high-income countries where such welfare support is widely
available, depression and other common mental disorders account for a signifi cant
proportion of overall payments.5 Instead, the analysis focused on the fi nancial benefi ts
fl owing from increased rates of workforce participation and productivity. The analysis only
considers the contribution to the economy as a whole through increased economic output;
it does not estimate the various income shares of this output.
Social eff ects
Conceptually distinct from improvements in clinical functioning (health eff ect) and the
restored ability to do paid work (economic eff ect), the successful treatment of depression
and anxiety disorders leads to improved opportunities for individuals and households to
pursue their leisure interests, participate more in social and community activities, and
carry out household production roles. The economic worth of these non-market
production and welfare gains is incorporated into our estimate of the intrinsic value of
mental health.
Articles
418
www.thelancet.com/psychiatry Vol 3 May 2016
integrated model of chronic disease management. These
include programme management and administration,
training and supervision, drug safety monitoring, health
promotion and awareness campaigns, and strengthened
logistics and information systems. We expressed
estimates as an on-cost to the estimated direct health-
care costs. The baseline value for this on-cost was 10%
(and therefore grows in absolute terms during scale-up).
Intervention coverage
The appendix provides coverage rates used for each
individual intervention at diff erent levels of national
income (appendix p 3). Summing across all interventions
and their respective populations in need, it is estimated
that—depending on the income level of the country—
between 7% and 28% of all people with depression
currently receive treatment, equivalent to a treatment gap
of 72–93% (table 1). A gradual, linear increase in treat-
ment coverage to a third of all cases in low-income
countries and to more than half of cases in high-income
countries would close the current gap by 29–39%; the use
of separate target coverage rates for low-income, middle-
income, and high-income countries refl ects diff erences
in which they stand now with respect to treatment
coverage, and are intended to refl ect what has been
achieved through programme scale-up eff orts in countries
such as Chile and the UK.20 Because of even lower starting
coverage levels, the modelled gap reduction for anxiety
disorders is lower than for depression (16–25%).
Eff ect of labour force on treatment
We modelled the economic eff ect of decreased morbidity
in terms of increased participation in and increased
productivity of the workforce. With regards to labour
force participation, very few studies have assessed the
extent to which eff ective depression treatments get
people back into work, and when measured, estimates
have been subject to local factors such as prevailing levels
of unemployment in the economy (panel 2).21–24 For our
base case, we conservatively modelled a 5% restored
ability to work as a result of treatment, with half and
double that rate used under pessimistic and optimistic
scenario analyses. Impaired productivity was assessed
both with respect to whole days off work (absenteeism)
and also partial days of impaired activity while an
individual is at work (presenteeism). Compared with
adults without common mental disorders in a range of
low-income, middle-income, and high-income countries
participating in the World Mental Health Survey,
4–15 more days out of role per year were recorded
because of depression and 8–24 days because of
generalised anxiety disorders; additional time lost per
year due to presenteeism was 11–25 partial disability days
for depression and 12–26 for generalised anxiety
disorders.25,26 Again, there are few empirical studies upon
which to base estimates of the eff ect of eff ective treatment
of depression and anxiety on productivity, and these
point towards small diff erences between intervention
and control groups (panel 2).27–33 Expressed as a proportion
of total working days per year (220 days), and allowing
for both the onset of eff ect as well as the time lag between
improved health and return to work, we modelled a 5%
increase in working days as a result of reduced
absenteeism, and a 5% increase through reduced
presenteeism. Again, these baseline values were varied
up and down by a factor of 2 and 0·5, respectively, in an
uncertainty analysis. These losses in and returns to
productivity were linked to the prevailing rates of labour
participation in the working age population (age
15–65 years) and gross domestic product (GDP) per
worker in each of the 36 assessed countries34,35 to calculate
productivity losses at current levels of treatment coverage
and productivity gains after scaled-up treatment. The
model does not account for potential changes in
retirement age or working patterns over time, although
an increase in retirement age and more fl exible working
patterns might enhance the overall productivity gains by
people with depression and anxiety with treatment.
Economic value of health benefi ts
Improvements in labour force outcomes represent the
instrumental value of improved mental health after
eff ective treatment of common mental disorders.
Independent of this instrumental value, being alive and
healthy is also valuable in itself. For this analysis, we
followed the approach adopted by Stenberg and
colleagues,36 who divided the overall value of a life-year
into its economic (instrumental) and health (intrinsic)
elements. For the Lancet Commission on Investing in
health, the value of a 1 year increase in life expectancy in
low-income and middle-income countries was estimated
to be 2·3 times per person national income, and
1·6 times per person national income worldwide (using a
discount rate of 3%).7 Stenberg and colleagues36 attributed
two-thirds of that derived value to the instrumental
components, which are measured here directly via the
Current
coverage
Target
coverage
Current gap Reduced gap % gap
reduction
Depression
Low-income countries 7% 34% 93% 66% 29%
Lower middle-income countries 14% 42% 86% 58% 32%
Upper middle-income countries 21% 49% 79% 51% 35%
High-income countries 28% 56% 72% 44% 39%
Anxiety disorders
Low-income countries 5% 20% 95% 80% 16%
Lower middle-income countries 10% 30% 90% 70% 22%
Upper middle-income countries 15% 35% 85% 65% 24%
High-income countries 20% 40% 80% 60% 25%
*Treatment coverage was modelled to increase from current to target rates linearly.
Table 1: Current and target levels of scaled-up treatment coverage for depression and anxiety disorders
(all interventions combined), by country income level*
Articles
www.thelancet.com/psychiatry Vol 3 May 2016
419
labour force outcomes, leaving the remaining third for
the intrinsic benefi ts of health, which is equivalent to
0·5 times per person income.
Uncertainty analysis
We assessed the sensitivity of results to plausible
variations around these and other key input parameters
by constructing optimistic and pessimistic scale-up
scenarios. For the upper estimate: total investment costs
were assumed to be 20% lower than baseline, as a result
of lower than expected use of expensive hospital
outpatient and inpatient care or the development of more
effi cient interventions, including internet-based treat-
ments; and productivity eff ects were set at double their
baseline rate (10% rather than 5%); the intrinsic value of
a year of health life was set at 0·7 times GDP per person
(rather than 0·5). For the lower estimate: total investment
costs were assumed to be 20% higher than baseline, as a
result of higher than expected drug prices, service use
and programme management; productivity eff ects were
set at half their baseline rate (2·5% rather than 5%); and
the intrinsic value of a year of health life was set at
0·3 times GDP per person (rather than 0·5).
Role of the funding source
The funders of the study had no role in study design,
data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had full
access to all the data in the study and had fi nal
responsibility for the decision to submit for publication.
Results
Across the 36 largest countries in the world, in the absence
of scaled-up treatment, it is projected that more than
12 billion days of lost productivity (equivalent to more than
50 million years of work) are attributable to depression and
anxiety disorders every year, at an estimated cost of
US$925 billion. Assuming the same distribution of costs
across lower-income and higher-income countries holds
for all other countries (representing 20% of the world’s
population), the global cost per year is $1·15 trillion.
Compared with people without these disorders, 4·7 billion
extra days are lost, at a cost of $592 billion (36% of the total
cost); this fi gure can be termed the excess productivity loss
of these disorders (fi gure 1).
Table 2 shows the estimated cost of scaling up treatment
for depression and anxiety, expressed as the net present
value of the total expenditure required over the scaling-up
period between 2016 and 2030 (ie, the cumulative cost
over 15 years of steady scale-up, but discounted at a rate of
3%). These costs relate to incremental treatment coverage
in the population over and above current levels of
coverage. For all 36 countries, the total cost amounts to
US$91 billion for depression and $56 billion for anxiety
disorders. Treatment of mild cases accounts for less than
10% of total costs for depression and 20% for anxiety
Panel 2: Labour force eff ects of treatment
Labour force participation
There are very few studies showing the extent to which eff ective
depression treatments get people back into work. Two studies
undertaken in the USA reported a 6% increase in employment
retention in patients with depression whose care was monitored
and managed closely.20,21 Findings of another US study22 of patients
in primary care showed that, at 6 months, employment rates were
52·5% for patients with no care versus 72·2% for patients with care.
For low-income and middle-income countries, programme
evaluation data for livelihoods from four countries—China, India,
Ghana, and Pakistan—were made available by BasicNeeds, which
showed that the proportion of people with depression undertaking
income-generating activities increased by more than 50%, and in
those with anxiety by more than 30% (Chris Underhill, BasicNeeds,
personal communication). These estimates are in line with the
assessment of the BasicNeeds programme in Kenya, which for a
more mixed caseload showed an 43% improvement in the
proportion of enrollees in income generation or productive work.23
Because these data are based on observation rather than under
controlled trial conditions, we can infer only a clear association
between exposure to treatment and subsequent earnings rather
than a defi nitive eff ect of intervention. For our base case, we
therefore conservatively modelled a 5% restored ability to work as
a result of treatment, with half and double that rate used under
pessimistic and optimistic scenario analyses.
Labour force productivity
A comprehensive review of 440 published trials in an existing
database of psychological and pharmaceutical interventions in
depression24 was specifi cally undertaken for this project (by
researchers at the Vrije Universiteit Amsterdam, Amsterdam,
Netherlands, and the Trimbos Institute, Utrecht, Netherlands)
to identify the eff ect of eff ective treatment on productivity;
unfortunately, very few trials reported these eff ects. However,
some treatment trials done in the USA, Korea, and India have
estimated the eff ect of intervention on productivity loss. The
decrease in absenteeism reported in these studies was close to
1 day per month.20,29–32 Only two studies reported the fi ndings
for presenteeism separately from days lost because of
absenteeism: in the Korean study, treated patients had 24 more
productive hours per month,29 whereas in the Indian study,
patients receiving the collaborative care had 4 fewer partial
days lost than controls.30 By conservatively assuming that
1 partial day is equivalent to a third of a whole day, we estimate
that almost 1 complete day of unimpaired work is restored per
month through reduced presenteeism. Expressed as a
proportion of total working days per year (220 days), and
allowing for both the onset of eff ect and the time lag between
improved health and return to work, a 5% increase in working
days is gained through reduced absenteeism, and a 5% increase
through reduced presenteeism.
For more on BasicNeeds see
http://www.basicneeds.org
Articles
420
www.thelancet.com/psychiatry Vol 3 May 2016
disorders. After standardising for population size, the
cost is actually quite low; for depression treatment, the
average annual cost during 15 years of scaled-up
investment is $0·08 per person in low-income countries,
$0·34 in lower middle-income countries, $1·12 in upper
middle-income countries and $3·89 in high-income
countries (table 2). Per person costs for anxiety disorders
are nearly half that of depression.
Table 2 shows results for two key health outcomes: cases
averted (reduced prevalence) and healthy life-years gained
(equivalent to disability-adjusted life-years averted). Across
the 36 countries represented in the analysis, we recorded a
small decrease in the estimated prevalence of depression
and anxiety disorders as a result of treated cases recovering
from illness more quickly; in the next 15 years, this gradual
decrease in prevalence translates into millions of averted
cases (73 million fewer cases of depression, and 45 million
fewer cases of anxiety disorder). Weighting these averted
prevalent cases by the average level of improved
functioning or reduced disability provides a measure of
healthy life-years gained. For depression and anxiety
disorders combined, the cumulative number of healthy
life-years gained over 15 years is 43 million.
Table 2 also shows the diff erence in aggregate GDP
between a continued current coverage scenario and one
refl ecting scaled-up treatment and enhanced productivity;
again, this and the total economic return for the entire
period of scale-up has been discounted at 3% to give a net
present value. For all 36 countries combined, the net
present value is $399 billion ($230 billion for depression
and $169 billion for anxiety disorders). The intrinsic
value of health returns show a net present value of more
than $250 billion for scaled-up depression treatment and
more than $50 billion for anxiety disorders (table 2).
Low-income
countries (N=6)
Lower
middle-income
countries (N=10)
Upper
middle-income
countries (N=10)
High-income
countries (N=10)
All countries
(N=36)
Total population of countries analysed (millions, 2013) 443 2215 2101 992 5751
Depression
Total investment (net present value, US$ million) 517 7164 20 338 63 503 91 522
Average annual investment (net present value, US$
per person)
0·08 0·34 1·12 3·89 1·50
Health returns (averted prevalent cases) 6 150 311 25 989 404 25 607 740 15 750 268 73 497 723
Health returns (healthy life-years gained) 2 234 781 15 692 290 11 414 429 7 567 211 36 908 711
Economic returns (US$ millions) 1190 18 799 52 732 157 022 229 744
Value of health returns (US$ millions)* 991 21 679 56 435 178 588 257 694
Benefi t to cost ratio (economic returns) 2·3 2·6 2·6 2 ·5 2· 5
Benefi t cost ratio (economic and value of health
returns)
4·2 5·7 5·4 5·3 5·3
Anxiety disorders
Total investment (net present value, US$ millions) 30 4 3797 8966 42 668 55 735
Average annual investment (net present value, US$
per person)
0·05 0·16 0·52 2·44 0·88
Health returns (averted prevalent case) 3 395 363 16 59 719 12 980 180 12 077 053 45 052 316
Health returns (healthy life-years gained) 416 232 2 220 716 1 711 767 1 604 069 5 952 783
Economic returns (US$ millions) 824 11 578 26 691 129 705 168 797
Value of health returns (US$ millions)* 18 1 2966 8453 40 409 52 009
Benefi t cost ratio (economic returns) 2·7 3·0 3·0 3· 0 3·0
Benefi t cost ratio (economic and value of health
returns)
3·3 3·8 3·9 4·0 4·0
*Healthy life-years gained multiplied by GDP per person multiplied by 0·5.
Table 2: Costs and benefi ts of scaled up treatment of depression and anxiety disorders, 2016–30
Figure 1: Lost productivity attributable to depression and anxiety disorders
at current treatment coverage, by country income level (US$ billion, 2013)
0
$50
$100
$150
$200
$250
$300
$350
$400
$450
$500
Productivity loss (US$ billion, 2013)
Low-income and middle-
income countries (n=26)
High-income
countries (n=10)
Excess loss
Comparative loss
$185
$276
$148
$316
Excess loss Comparative loss
Articles
www.thelancet.com/psychiatry Vol 3 May 2016
421
By summing the discounted costs and benefi ts for all
countries in an income group, we derived a summary
measure of the relationship between the benefi ts of
scaled-up treatment and the associated costs of
investment (table 2, fi gure 2). Restricting assessment to
the economic returns to investment, benefi t to cost
ratios for scaled-up depression treatment across
country income groupings were in the range of
2·3 to 2·6. For anxiety disorders the ratios were slightly
higher (range 2·7–3·0). Extension of the benefi t–cost
analysis to include the estimated value of health returns
increased the ratio of benefi t to cost, especially for
depression because of the higher health returns for this
disorder compared with anxiety disorders. Benefi t to
cost ratios for depression now exceed those for anxiety
disorders (range 4·2–5·7), and were more than double
the ratio when only economic benefi ts of depression
treatment scale-up were considered. Benefi t to cost
ratios for anxiety disorders increased by a third
(range 3·3–4·0).
We did uncertainty analysis to ascertain the sensitivity
of results to plausible changes in key study parameters.
Benefi t to cost ratios fell to or almost reached parity
under the more pessimistic scenario when only economic
benefi ts were considered, and did not exceed 3 even
when the value of health benefi ts was included (fi gure 2).
By contrast, the more optimistic scenario produces
benefi t to cost ratios of 5·5–7·2 (economic benefi ts only)
and 7·5–11·3 when the value of health benefi ts was
added in. As expected, results were quite sensitive to the
estimated rate of enhanced labour participation and
productivity. We also assessed the eff ect of changing the
rate used to discount future costs and benefi ts to the
present time. At a discount rate of 6%, the net present
value of total investments and returns would be 25%
less; with no discounting, they would be 35% higher in
absolute terms. Because such a change in discount rate
was applied to both costs and benefi ts, the ratio of benefi t
to cost, our summary return on investment metric, is not
aff ected.
Discussion
This analysis sets out, for the fi rst time, a global
investment case for a scaled-up response to the massive
public health and economic burden of depression and
anxiety disorders. Previous international economic
studies of mental health have assessed the economic
eff ect of these disorders,2,3 the cost-eff ectiveness of
diff erent intervention strategies,8,10 and the cost of scaling
up care,18,19 but not the value of both economic and health
benefi ts of intervention scale up.
Notwithstanding the general limitations of any
projection modelling study, the analysis suggests that the
investment needed to substantially scale up eff ective
treatment coverage for depression and anxiety disorders
in the 36 countries included in this analysis is substantial;
the net present value of all investments between 2016
and 2030 is $147 billion, equivalent to less than $10 billion
per year on average. Extending the scope to the 20% of
the world’s population not living in the 36 countries
represented in the study would increase the cost by about
25% to $184 billion. However, the returns to this
Figure 2: Baseline, upper, and lower benefi t to cost ratios for scaled-up treatment of depression and anxiety
disorders, by country income group
0
2
4
6
8
10
12
Benefit:cost ratio
0
2
4
6
8
10
12
Benefit:cost ratio
0
2
4
6
8
10
12
Benefit:cost ratio
0
2
4
6
8
10
12
Benefit:cost ratio
Economic benefits: depression
Economic and value of health benefits: depression
Economic benefits: anxiety
Economic and value of health benefits: anxiety
Low-income
countries (n=6)
Lower middle-
income countries
(n=10)
Upper middle-
income countries
(n=10)
High-income
countries (n=10)
5·5 6·0 6·2 5·9
2·3 2·6 2·6 2·5
1·0 1·1 1·0 1·0
6·4 7·0 7·2 7·2
2·7 3·0 3·0 3·0
1·0 1·1 1·0 1·0
8·8
11·3 11·1 10·9
4·2
5·7 5·4 5·3
1·9
3·0
2·2 2·6
7·5
8·3 8·9 8·9
3·3 3·8 3·9 4·0
1·5 1·7 1·6 1·8
Articles
422
www.thelancet.com/psychiatry Vol 3 May 2016
investment are also substantial, with benefi t to cost ratios
of 2·3–3·0 when economic benefi ts only are considered,
and 3·3–5·7 when the value of health returns are also
included. To put these fi ndings into context, any benefi t
to cost ratio exceeding 1 provides a rationale for
investment. Compared with some other potential
investments in health, ratios of the order reported here
can be deemed relatively modest. For example, a return
on investment analysis for malaria, also for 2016–30, but
using the full value of a statistical life-year, estimated
benefi t to cost ratios in the range of 28:1 to 40:1.37 An
investment case done for maternal, reproductive,
neonatal, and child health obtained a benefi t to cost ratio
of less than 10:1 for 2013–35,36 which is closer to the
results obtained in this study. Inclusion of other benefi ts
arising from scaled-up treatment of common mental
disorders that could not be captured though the present
modelling exercise, notably reduced welfare support
payments, and improved outcomes for other aff ected
people (eg, partners and children of women with
perinatal depression) would generate higher ratios of
benefi t to cost. Set against that, treatment programmes
might cost more or achieve less than anticipated, as
highlighted by the uncertainty analysis.
One limitation of our study is that although the
projected level of overall prevalence of depression and
anxiety disorders is quite well-established,12-14 the same
cannot be said for treated prevalence. The analysis done
here allows for a gradual linear increase in eff ective
service coverage for depression and anxiety disorders in
all parts of the world in the next 15 years. However, for
this to happen, not only will a new level of political
commitment and resource mobilisation be required, but
also a signifi cant reorientation of public health systems
towards chronic disease identifi cation and management.9
Partial or weak implementation of envisaged treatment
programmes, including appropriate management of
recurrent cases of depression or insuffi cient promotion
and awareness programmes, will inevitably reduce the
number of cases eff ectively reached and therefore the
health and other benefi ts obtained. It is also possible that
as treatment coverage in the population increases
substantially, the average cost per case might go up, for
example as a result of reaching out to more remote or
less well-served parts of a country. Target coverage rates
were accordingly set at a modest level in this analysis (an
upper value of 56% of depression cases in high-income
countries). Aside from projected treatment coverage and
eff ectiveness, a further crucial parameter for this analysis
concerns the eff ect of treatment on labour force
participation and productivity, for which there remains a
paucity of evidence. As concluded by a systematic review,
such data are not hard to collect alongside clinical trials
and other studies, and need to be uniformly measured
more often.27 More generally, population health models
(eg, the OneHealth tool) rely on many input parameters,
data sources, and assumptions regarding expected rates
of disease, demographic change, and intervention eff ects
in the future, which limits their precision.
Several eff ects were not included in the analysis. One
was the negative eff ect of maternal depression on early
child development, for which there is clear evidence;38
the health, social, and economic benefi ts of eff ective
treatment of maternal depression on the cognitive and
physical development of newly born babies was not
assessed, but there is some evidence that this could be
substantial over the longer term.39 Likewise, the monetary
and non-monetary impact of eff ective treatment on
family and other caregivers has not been factored in.
Additionally, no account has been taken of the substantial
eff ect of depression and its treatment on physical health
outcomes; depression is a risk factor for disorders such
as hypertension, stroke, coronary heart disease, and
substance use disorders (just as these conditions are risk
factors for depression), and adversely aff ects outcomes
through reduced help-seeking and adherence.40 Inclusion
of these additional eff ects of treatment would bolster
identifi ed economic returns. Taking appropriate account
of the regular co-occurrence of depression and anxiety in
individuals would be expected to lead to strong synergies
on the treatment side, leading to potentially reduced
investment costs, but health and economic outcomes for
these comorbid cases might be slower or harder to
achieve.
Although the analysis accounted for age and sex (eg, in
terms of disease prevalence, labour force participation
and treatment eligibility), it was not possible to consider
the eff ect of socioeconomic status as a mediator and
predictor of good health and economic outcomes.
Poverty has an adverse eff ect on the risk of depression
and anxiety disorders through higher levels of stress,
social exclusion, violence and trauma, but the evidence
base for the mental health eff ect of interventions
targeted at the poor remains insubstantial.41 In many
countries, poor people face signifi cant barriers to
accessing services, including the fi nancial cost of
seeking and paying towards health care. Finally, it should
be acknowledged that the workplace itself can be a
source of stress for many people, and that there is a
consequent need to integrate mental health and
wellbeing into new or existing employee support
programmes.
A crucial issue related to but outside the scope of this
return on investment analysis is the source of fi nancing
for investments required to scale-up services for
depression and anxiety disorders. As previously noted,
the absolute amount needed for investment (eg, on a per
person basis) is modest, but because existing service
coverage level is so low in most countries, the gap
between current and required spending can be large.18,19
Accordingly, both rich and poor countries need to
carefully consider the merits of diff erent health fi nancing
mechanisms. For many countries, the fi rst question to
address concerns the extent to which domestic fi nancing
Articles
www.thelancet.com/psychiatry Vol 3 May 2016
423
represents a feasible and suffi cient method for fi nancing
mental health services, perhaps as part of a package of
measures to be paid for from enhanced revenue
generation. For low-income countries eligible for offi cial
development assistance, a second question might be to
what extent external funding can complement
domestically generated resources to catalyse service
development. In countries where domestic or external
funding mechanisms are expected to fall short of
requirements or pose a risk to fi scal stability, a further
question relates to the extent to which market-based
fi nancing options such as bonds off er a suitable and
feasible approach to generating and providing funds for
outcomes-based scale-up for mental health services.
The pursuit of any of these methods of fi nancing will
be aff ected by other factors, including the amount of
investment needed, the level of political will and also
fi scal space for raising new resources for health, and
eligibility of the country for bilateral or multilateral
funding. Faced with a new and broad development
agenda,6 governments need to assure themselves that
investment in the mental health of their populations
represents a sound and equitable investment of society’s
resources that leads to clear and defi nable health,
economic, and social benefi ts. Our return on investment
analysis, coupled with an assessment of health-system
needs and priorities, and the broader macro-fi scal
situation, can contribute to a balanced investment case
for common mental disorders and the health sector more
generally.
Contributors
DC and SS conceived, planned, and oversaw the study. DC led the
analysis of treatment costs and health outcomes, and drafted the paper.
KS led the development of the methodology and model for estimating
productivity eff ects. PS and BR contributed to the conceptual
development of the return on investment model and its constituent parts.
PC and FS led the systematic review of productivity eff ects of treatment.
All authors reviewed, commented on, and approved the report.
Declaration of interests
We declare no competing interests.
Acknowledgments
We thank the contribution of Chris Underhill, Jess McQuail, and
Uma Sunder of BasicNeeds, who provided country-level data for
productivity outcomes for people enrolled into their mental health and
development programme; colleagues in the Department of Health
System Governance and Financing (Jeremy A Lauer and
Melanie Bertram) regarding conceptualisation and review of return on
investment frameworks in health, and Eirini Karyotaki at VU
Amsterdam University for her contribution to a systematic review;
Shelly Chopra for her contribution to a background literature search and
to the development of the conceptual framework used in this study .
Development of the mental health module of the OneHealth tool, which
was used in this analysis for estimating health impacts, was made
possible through the EMERALD project on mental health system
strengthening in low-income and middle-income countries, which is
funded by the European Union under the 7th Framework programme
(Grant agreement 305968). DC and SS are staff members of WHO. All
opinions expressed in this report rest with the authors; they do not
necessarily represent the decisions, policy, or views of WHO.
References
1 WHO. Mental health ATLAS 2014. Geneva: World Health
Organization, 2015.
2 Bloom DE, Cafi ero E, Jané-Llopis E, et al. The global economic
burden of noncommunicable diseases. Geneva: World Economic
Forum, 2011.
3 Hu TW. Perspectives: an international review of the national cost
estimates of mental illness, 1990–2003. J Ment Health Policy Econ
2006; 9: 3–13.
4 Gustavsson A, Svensson M, Jacobi F, et al, and the CDBE2010 Study
Group. Cost of disorders of the brain in Europe 2010.
Eur Neuropsychopharmacol 2011; 21: 718–79.
5 OECD. Fit Mind, Fit Job: From Evidence to Practice in Mental
Health and Work. Paris: OECD, 2015.
6 United Nations. Transforming our world: the 2030 agenda for
sustainable development. New York: United Nations, 2015.
7 Jamison DT, Summers LH, Alleyne G, et al. Global health 2035:
a world converging within a generation. Lancet 2013; 382: 1898–955.
8 Hyman S, Chisholm D, Kessler R, Patel V, Whiteford H.
Mental disorders. In: Jamison DT, Breman JG, Measham AR, et al,
eds. Disease control priorities in developing countries, 2nd edn.
Oxford University Press and The World Bank, 2006: 605–26.
9 Patel V, Chisholm D, Parikh R, et al. Addressing the burden of
mental, neurological, and substance use disorders: key messages
from Disease Control Priorities, 3rd edn. Lancet 2015; published
online Oct 7, http://dx.doi.org/10.1016/S0140-6736(15)00390-6.
10 Chisholm D, Saxena S. Cost eff ectiveness of strategies to combat
neuropsychiatric conditions in sub-Saharan Africa and South East
Asia: mathematical modelling study. BMJ 2012; 344: e609.
11 United Nations Population Division 2015, World Population
Prospects, the 2015 Revision, at http://esa.un.org/unpd/wpp/.
12 Whiteford HA, Degenhardt L, Rehm J, et al. Global burden of
disease attributable to mental and substance use disorders: fi ndings
from the Global Burden of Disease Study 2010. Lancet 2013;
382: 1575–86.
13 Baxter AJ, Scott KM, Vos T, Whiteford HA. Global prevalence of
anxiety disorders: a systematic review and meta-regression.
Psychol Med 2013; 43: 897–910.
14 Ferrari AJ, Somerville AJ, Baxter AJ, et al. Global variation in the
prevalence and incidence of major depressive disorder: a systematic
review of the epidemiological literature. Psychol Med 2013; 43: 471–81.
15 van Zoonen K, Buntrock C, Ebert DD, et al. Preventing the onset of
major depressive disorder: a meta-analytic review of psychological
interventions. Int J Epidemiol 2014; 43: 318–29.
16 WHO. mhGAP (Mental Health Gap Action Programme)
Intervention Guide. Geneva: World Health Organization, 2010.
17 WHO-CHOICE. Country-specifi c unit costs. http://www.who.int/
choice/country/country_specifi c (accessed Nov 2, 2015).
18 Chisholm D, Lund C, Saxena S. Cost of scaling up mental
healthcare in low- and middle-income countries. Br J Psychiatry
2007; 191: 528–35.
19 Chisholm D, Burman-Roy S, Fekadu A, et al. Estimating the cost of
implementing district mental healthcare plans in fi ve low- and
middle-income countries: the PRIME study. Br J Psychiatry 2016;
208 (suppl 56): s71–78.
20 De Silva MJ, Lee L, Fuhr DC, et al. Estimating the coverage of
mental health programmes: a systematic review. Int J Epidemiol
2014; 43: 341–53.
21 Wang PS, Simon GE, Avorn J, et al. Telephone screening, outreach,
and care management for depressed workers and impact on clinical
and work productivity outcomes: a randomized controlled trial.
JAMA 2007; 298: 1401–11.
22 Wells KB, Sherbourne C, Schoenbaum M, et al. Impact of
disseminating quality improvement programs for depression in
managed primary care: a randomized controlled trial. JAMA 2000;
283: 212–20.
23 Schoenbaum M, Unützer J, McCaff rey D, Duan N, Sherbourne C,
Wells KB. The eff ects of primary care depression treatment on
patients’ clinical status and employment. Health Serv Res 2002;
37: 1145–58.
24 Lund C, Waruguru M, Kingori J, et al. Outcomes of the mental
health and development model in rural Kenya: a 2-year prospective
cohort intervention study. In Health 2013; 5: 43–50.
25 Alonso J, Petukhova M, Vilagut G, et al. Days out of role due to
common physical and mental conditions: results from the WHO
World Mental Health surveys. Mol Psychiatry 2011; 16: 1234–46.
Articles
424
www.thelancet.com/psychiatry Vol 3 May 2016
26 Bruff aerts R, Vilagut G, Demyttenaere K, et al. Role of common
mental and physical disorders in partial disability around the world.
Br J Psychiatry 2012; 200: 454–61.
27 Cuijpers P, van Straten A, Warmerdam L, Andersson G.
Psychological treatment of depression: a meta-analytic database of
randomized studies. BMC Psychiatry 2008; 8: 36.
28 Nieuwenhuijsen K, Faber B, Verbeek JH, et al. Interventions to
improve return to work in depressed people.
Cochrane Database Syst Rev 2014; 12: CD006237.
29 Harvey S, Joyce S, Modini M, et al. Work and depression/anxiety
disorders: a systematic review of reviews. 2012. Commissioned
review by University of South Wales for Beyond Blue. https://www.
beyondblue.org.au/docs/default-source/research-project-fi les/
bw0204.pdf?sfvrsn=4 (accessed April 16, 2015).
30 Woo J-M, Kim W, Hwang T-Y, et al. Impact of depression on work
productivity and its improvement after outpatient treatment with
antidepressants. Value Health 2011; 14: 475–82.
31 Buttorff C, Hock RS, Weiss HA, et al. Economic evaluation of a
task-shifting intervention for common mental disorders in India.
Bull World Health Organ 2012; 90: 813–21.
32 Rollman BL, Belnap BH, Mazumdar S, et al. A randomized trial to
improve the quality of treatment for panic and generalized anxiety
disorders in primary care. Arch Gen Psychiatry 2005; 62: 1332–41.
33 Rost K, Smith JL, Dickinson M. The eff ect of improving primary
care depression management on employee absenteeism and
productivity. A randomized trial. Med Care 2004; 42: 1202–10.
34 ILO. 2015 databases and subjects. http://www.ilo.org/global/
statistics-and-databases/lang--en/index.htm (accessed April 2015).
35 World Bank. World development indicators. 2015. http://databank.
worldbank.org/data/reports.aspx?source=world-development-
indicators (accessed April 13, 2015).
36 Stenberg K, Axelson H, Sheehan P, et al, and the Study Group for
the Global Investment Framework for Women’s Children’s Health.
Advancing social and economic development by investing in
women’s and children’s health: a new Global Investment
Framework. Lancet 2014; 383: 1333–54.
37 World Health Organization on behalf of the Roll Back Malaria
Partnership Secretariat. Action and Investment to defeat Malaria
2016–2030: For a Malaria-Free World. 2015. World Health
Organization, Geneva.
38 Rahman A, Fisher J, Bower P, et al. Interventions for common
perinatal mental disorders in women in low- and middle-income
countries: a systematic review and meta-analysis.
Bull World Health Organ 2013; 91: 593–601.
39 Cuijpers P, Weitz E, Karyotaki E, Garber J, Andersson G. The eff ects
of psychological treatment of maternal depression on children and
parental functioning: a meta-analysis. Eur Child Adolesc Psychiatry
2015; 24: 237–45. DOI:10.1007/s00787-014-0660-6.
40 Prince M, Patel V, Saxena S, et al. No health without mental health.
Lancet 2007; 370: 859–77.
41 Lund C, De Silva M, Plagerson S, et al. Poverty and mental
disorders: breaking the cycle in low-income and middle-income
countries. Lancet 2011; 378: 1502–14.