1. health and fitness
2. disease
3. aids and hiv

Supplementary appendix
This appendix formed part of the original submission and has been peer reviewed.
We post it as supplied by the authors.
Supplement to: Degenhardt L, Whiteford HA, Ferrari AJ, et al. Global burden of disease
attributable to illicit drug use and dependence: findings from the Global Burden of
Disease Study 2010. Lancet 2013; published online Aug 29. http://dx.doi.org/10.1016/
S0140-6736(13)61530-5.
Disability weights
Disability was defined as any short or long-term health loss resulting from a disorder. This was calculated (in the form of disability weights) for each form of drug
dependence and multiplied by the prevalence data to derive YLDs1. In response to criticism2 of the reliability of earlier estimates of disability weights3, GBD 2010 calculated
new disability weights collected as wide a range of estimates as possible1 by using 1) representative community surveys in Bangladesh, Indonesia, Peru, the United Republic
of Tanzania and the United States of America (total n = 30,000) and 2) an internet-based survey. Survey participants were shown a number of randomly generated pair-wise
comparisons of different health states and asked to choose the ‘healthier’ of the pair. To estimate disability weights, responses were transformed into discrete values and
anchored between 0 (perfect health) and 1 (death) using some additional population health equivalence questions1.
For each illicit form of drug dependence, the proportion of asymptomatic cases were estimated from the US National Epidemiological Survey on Alcohol and Related
Conditions (NESARC) 2000-2001 and 2004-20054 and the Australian National Survey of Mental Health and Wellbeing of Adults (AHS) 19975. These proportions were used
to calculate an average disability weight for each disorder, where the proportion of asymptomatic cases was given a disability weight of 01. Further details are reported in1, 6..
Estimates are presented in table A1 below.
Table A1: Details of drug dependence disability weights
Cannabis dependence
Amphetamine dependence
Cocaine dependence
Opioid dependence
0.329 (0.223-0.455)
0.353 (0.215-0.525)
0.376 (0.235-0.553)
0.641 (0.459-0.803)
Asymptomatic
51% (47-54%)
45% (37-52%)
45% (37-52%)
16% (11-22%)
Symptomatic
49% (46-53%)
55% (48-63%)
55% (48-63%)
84% (78-89%)
0.162 (0.109-0.224)
0.193 (0.117-0.296)
0.209 (0.128-0.313)
0.501 (0.330-0.689)
Unadjusted disability weight
Case
Cases without disability (comorbidity adjusted)
Average disability weight (comorbidity adjusted)
2 Figure A1: The global coverage and number of data points obtained for each drug from our systematic review of the epidemiological
literature.
Note. Map shows the global coverage of prevalence, incidence, remission, and/or excess-mortality data included in the DisMod-MR modelling process; CA: cannabis
use/dependence, Co: cocaine dependence, AM: amphetamine dependence, Op: opioid dependence.
3 Modelling death estimation
Premature mortality due to illicit drug dependence was computed as years of life lost (YLLs) based on cause of
death estimates from 1980 to 2010 for 20 age-groups, for both sexes, in 187 countries (for full details see7).
Mortality was explicitly modelled for illicit drug use disorders. Cause of death estimates were developed using a
comprehensive database of vital registration, verbal autopsy, surveillance, and other sources. Ultimately, 9,075
country-years of data from 124 countries were used for the drug use disorders mortality estimation. The quality
of each observation was assessed, and various revisions of ICD classifications mapped. Deaths were reassigned
using standardised algorithms where the recorded cause of death was not likely to be the underlying cause of
death. Deaths attributable to drug dependence are often mischaracterised as accidental poisonings. Deaths coded
as accidental poisonings due to drugs that fell under the drug-use disorder category included narcotics,
hallucinogens, sedative-hypnotic or psychotropic drugs. These were recoded to drug use disorder deaths, unless
they occurred in children.
An ensemble modelling strategy was used for death estimation, employing mixed effects linear models and
spatiotemporal Gaussian process regression models weighted by out-of-sample predictive validity. This process
is described in detail elsewhere.8, 9 Where data were sparse or missing, the models were informed by borrowing
strength across space and time and by using relevant covariates. All models included a transformed measure of
average income per capita, education level, and an aggregate measure of health system access, which takes into
account indicators such as hospital beds per capita, in facility deliveries, and vaccination coverage rates. Drug
use disorder models were further informed both by indicators of use, such as opium production, and by risk
factors of cardiovascular disease that would make death due to overdose more likely, such as blood pressure.
In order to attribute drug use disorders to specific categories of use, the fraction attributable to each specific
category was extracted from vital registration and verbal autopsy data, pooled by year and region, and rescaled
to sum to one. The resulting year-region specific fractions were applied to the estimates of overall drug use
mortality to get estimates for opioid, cocaine, amphetamine and other drug related deaths. Uncertainty in cause
of death model predictions was captured using standard simulation methods by taking 1,000 draws for each age,
sex, country, year, and cause. To create consistency between the sum of cause-specific and all-cause mortality,
the models were re-scaled according to the uncertainty around the cause-specific rate. The resulting predicted
numbers of deaths were converted to the measure YLLs by multiplying deaths by the remaining life expectancy
at the age of death as derived from the GBD 2010 standard model life table. For full details of mortality
modelling see7.
4 Comparative risk assessment: cannabis use as a risk factor for schizophrenia
We considered several ways in which cannabis and schizophrenia may be linked: Model 1) a model that
assumed greater disorder severity among those using cannabis regularly who have already developed the
disorder; Model 2) a model that assumed the association reflects earlier onset of schizophrenia among those who
would have developed it anyway; Model 3) a model that assumed reduced remission from schizophrenia once it
has developed; and Model 4) a model that assumed increased incidence of schizophrenia.
After consideration, approaches 3 and 4 were excluded from core GBD analyses because of the lack of data
needed to systematically quantify the relationship across different studies while also accounting for confounding
variables. Approaches 1 and 2 were also deemed more plausible on the basis of the literature10.
Data on regular (weekly or more frequent) cannabis use in the past year
We found seven studies reporting prevalence of weekly or more frequent cannabis use in the past year, from 15
countries and five GBD world regions; and 80 studies on the prevalence of past-year cannabis use, from 82
countries and 17 GBD world regions.
The epidemiological data available for regular cannabis use were modelled using an age-integrating Bayesian
hierarchical model (DisMod-MR). Estimates of prevalence were derived separately for 21 world regions, males
and females, 5 year age groups. We assumed zero incidence and prevalence of regular cannabis use before age
13 as this led to the most plausible fit to the data. A study-level covariate was used to adjust estimates of past
year cannabis use towards the desirable which were estimates of weekly cannabis use. Prevalence from past year
use were 3.79 (3.48-4.13) times higher than estimates of weekly cannabis use and were adjusted downwards
accordingly.
The final models used are presented below in Figure A2.
5 Figure A2: Estimated regional distribution of regular cannabis use by sex, 2010
6 Modelling Earlier age of onset
The effect of cannabis use on schizophrenia was modelled via two pathways: the first is by causing the average
age of onset to be earlier than with no cannabis use, and the second is by increasing the severity of
schizophrenia. To account for the effects of the first pathway, we calculated the counterfactual average duration
of schizophrenia across all ages under a scenario of no cannabis use, and compared this to the currently
observed duration. To determine the counterfactual average duration of schizophrenia, we shifted incident cases
of schizophrenia who use cannabis to be 2.70 (95% CIs: 1.96-3.43) years later based on a systematic metaanalysis by Large et al11.
We used the estimates of cannabis use by age (in single years), sex, country and year described above, assuming
that the prevalence of regular cannabis was the same among individuals with schizophrenia and individuals
without schizophrenia. Estimates of the number of incident cases and the corresponding duration of
schizophrenia by age, sex, country and year were based on the DisMod model for Schizophrenia. The value of
one minus the ratio of the counterfactual to the observed duration is an estimate of the PAF of schizophrenia due
to the effect of regular cannabis use on age of onset.
Modelling Increased severity of schizophrenia
To calculate the burden associated with second pathway of shifting severity, we used the odds ratio from Foti et
al12 of psychotic symptoms of 1.64 (95% CIs: 1.12-2.34) in people with schizophrenia who regularly use
cannabis as opposed to those who do not. We converted the ORs to their RR equivalents based on the
prevalence of exposure to regular cannabis use and the outcome of percent of time with psychosis (as opposed
to residual state). The percent of time spent in acute psychosis was 63% (38%-82%) based on a meta-analysis of
6 studies covering 5 GBD world regions. We used the linear relationship between the estimated change in
disability weight (based on the proportion of time spent in a psychotic state) and the prevalence of regular
cannabis use to calculate the percent of schizophrenia disability attributable to regular cannabis use.
7 Comparative risk assessment: Opioid, amphetamine and cocaine dependence as risk
factors for suicide
A systematic literature review was conducted to identify studies reporting the relative-risk (RR) (or provided
enough information to calculate RRs) of opioid dependence, amphetamine dependence and cocaine dependence
as independent risk factors for suicide13. Studies were included if they adhered to DSM/ICD diagnostic criteria
and used prospectively collected data from representatives samples of the community.
We found 21 cohort studies for opioid dependence, three studies for cocaine dependence, and one study for
amphetamine dependence. For each drug type, a meta-analysis, using a random effects design, was used to pool
RR estimates. The pooled RRs have been summarised in table #. Given the lack of sex-specific data, the metaanalyses combined male and female estimates. To address the lack of RR data for amphetamine and cocaine
dependence, RR estimates for these disorders were pooled to derive an overall a psychostimulant dependence
RR estimate.
Pooled RRs were used together with disorder-specific exposure data (i.e. DisMod-MR prevalence estimates by
sex, age, year, and region) to calculate population attributable fractions (PAFs) for each disorder. PAFs were
then multiplied by suicide deaths and DALYs to calculate the proportion of suicide burden (from the GBD 2010
injuries cause group) attributable to each drug use disorder14.
Table A2: Pooled relative risks (RR) for suicide among people who are opioid, cocaine and amphetamine
dependent
Disorders
Number of studies
Pooled RR
Opioid dependence
21
6.9 (4.5-10.5)
Amphetamine dependence
3
8.2 (3.9-16.9)
Cocaine dependence
3
8.2 (3.9-16.9)
8 Comparative risk assessment: injecting drug use as a risk factor for HIV, HBV and
HCV
To estimate the population attributable fraction (PAF) for HIV, HBV and HCV, we first estimated the fraction
of HIV, HBV and HCV seroprevalence amongst people who inject drugs that is due to injecting drug use (IDU).
This fraction was determined using the relative risk of seroconversion for injecting drug use compared to no
injecting drug use from a published cohort study.15
Multiplying this fraction by the estimated prevalence of HIV, HBV and HCV amongst people who inject drugs
and then by the prevalence of IDU in the population gives the population-level prevalence of HIV, HBV and
HCV due to IDU. Dividing this quantity by the total population-level prevalence of HIV, HBV and HCV
provides an estimate of the population attributable fraction (PAF). This calculation was performed for adults
aged 15-64 years of age in each of the 187 GBD countries for 1990, 2005 and 2010. Formally, this is given as
1
where : c is cause (HIV, HBV, HCV)
i is country
t is time (1990, 2005, 2010)
RR is the relative risk of sero-conversion for injecting drug use compared to no injecting drug use
SPidu is sero-prevalence amongst IDUs
IDU is prevalence of IDU in the population
SPpop is sero-prevalence in the population
The assumed theoretical minimum risk exposure distribution is therefore zero. We used simulation analysis to
determine the uncertainty in PAF estimates.
Population-level HIV, HBV and HCV seroprevalence
Population-level HBV and HCV seroprevalence data were systematically collated with the majority of data
sources measuring levels of hepatitis C antibodies (anti-HCV) and hepatitis B surface antigen (HBsAg).
Estimates of HIV prevalence by country and year were based on 2012 UNAIDS estimates.5
As data were available for only selected country-time periods, we estimated population-level prevalence of HIV,
HBV and HCV in the general population for all country-time periods using an age-integrating Bayesian
hierarchical model (DisMod-MR). The model includes fixed effects for study-specific and national-level
covariates, and random effects by GBD super region, region, and country.
9 Population-level prevalence of injecting drug use
Data Sources
Data on injecting drug use was obtained through a multistage systematic search of peer-reviewed and grey
literature, and international consultation with experts in the IDU, HIV and hepatitis fields. Details of these
reviews have previously been reported2,3.
Briefly, a systematic search of multiple sources including MEDLINE, Embase, CAB abstracts, WHO library
(WHOLIST) and SIGLE (grey literature database) was conducted. Data sources were also identified and
collected using a standardized protocol to contact experts and request data. Additional studies were also
obtained through email contact with GBD Illicit Drug Use Expert Group members, UN agency staff, relevant
international email lists, and other team contacts. Reviews used estimates of current injectors defined as those
who had injected in the past 12 months as the optimal metric.
Quality of the data was assessed. Studies were also graded by quality (Grade A includes multisite
seroprevalence studies with multiple sample types. Grades B-E are of decreasing quality). The IDU population
was assumed to be people 15-64 years old and thus prevalence of IDU and prevalence of HIV, HBV and HCV
among IDUs was calculated among this population only.
When raw data were available, these were analysed directly. Where raw data was not available or was analysed
elsewhere, standardised protocols were used to ensure that the collection, analysis, and extraction of data were
done in a systematic and consistent manner. The final database included 178 data points (42 nationally
representative) from 59 countries.
Modelling Strategy
As data were available for only selected country-time periods, we estimated prevalence of IDUs for all countrytime periods using an age-integrating Bayesian hierarchical model (DisMod-MR). The model includes fixed
effects for study-specific and national-level covariates, and random effects by GBD super region, region, and
country. Estimates were producted for males and females in single year age groups, for 1990, 2005, and 2010.
Only prevalence data were available for modelling. We assumed zero incidence and prevalence before age 15 as
this led to the most plausible fit to the data. Heterogeneity was set at high across all parameters to reflect the
large variability in the dataset.
We used one study-specific covariate in the model to account for inconsistencies in the raw data. Data that were
sub-national (rather than nationally representative) are downweighted in the model through a study-specific
dummy variable for whether the data were sub-national or national. We also included age-standardised
prevalence rates by year from the opioid dependence model in GBD as a national covariate to improve the
prediction of prevalence of IDU. Table A3 below describes the study-level and national-level covariates
included in the model.
10 Table A3: Covariates included in model for injecting drug use
Covariates
Mean Effect Size
Effect Size SE
Study level covariate: sex
0.48
1.89
Study level covariate: sub-national data
National level covariate: age-standardised death rate from
opioid dependence
0.96
2.58
0.94
2.78
As demonstrated in the table, the opioid covariate had little effect, adjusting opioid dependence estimates
downwards towards IDU estimates by only 6%. The sub-national covariate, which created a crosswalk between
prevalence derived from a nationally representative sample (desirable) and prevalence derived from subnationally representative samples, adjusted sub-national samples upwards by a factor of 1.03 (1/0.96).
Many countries register people who inject drugs on government-maintained databases, for the purposes of
“tracking” and identifying people thought by police to be injecting drugs. These are sometimes the only source
of data on the size of the population of injectors in a given country. By necessity this estimate will also be an
underestimate, since not all injectors will be known to or registered by police.
Prevalence of HIV, HBV and HCV among injecting drug users
Data on the prevalence of HIV, HBV and HCV among IDUs were determined from two systematic reviews
where the strategy for collecting data is described in detail.16, 17 The process has also been described above in the
section of prevalence of IDU.
Nelson et al2 restricted the inclusion of hepatitis C reports to serological test results for hepatitis C antibodies
(anti-HCV). Studies for HBV were included if they reported serological testing for HBsAg or anti-HBc. If more
than one prevalence report met inclusion criteria, the lowest and highest reports were selected and the midpoint
between the two taken. If a report included lower and upper uncertainty bounds, these were used. If only one
prevalence report met inclusion criteria and did not report uncertainty bounds, regional or global limits were
applied to the country-level report. Mathers et al1 took the lowest lower bound and the highest upper bound as
the lower and upper bound for estimating HIV or IDU prevalence when two or more estimates for HIV or IDU
prevalence existed and ranges were given for both.
As estimates of sero-prevalence were reported for different years and were missing for several countries, we
used mixed effect regression to estimate sero-prevalence of HIV, HBV and HCV for all countries for 1990,
2005 and 2010. We used the medium estimated prevalence of HIV and HCV as the dependent variables of the
regression and HBV and HCV population-level prevalence, IDU population-level prevalence and the interaction
between the two as fixed effects. We included nested random effects for GBD super region, GBD region and
country. Simulation analysis was used to determine uncertainty in the estimated prevalence of HIV, HBV and
HCV among IDUs. The final models are presented below in Figure 2.
11 Relative risk of HIV, HBV or HCV seroconversion due to injecting drug use
Due to the inherent difficulties of conducting such studies, few epidemiological data are available to describe
relative risk of HIV, HBV and HCV seroconversion due to injecting drug use compared to people who do not
inject drugs. We estimated the relative risks due to injecting drug use using data from the Amsterdam Cohort
Study which described HIV and HCV seroconversion rates among different categories of drug dependence and
participation in a harm reduction program15.
To generate a relative risk, we compared seroconversion rates amongst individuals that were dependent and did
not participate in harm reduction with individuals who had no or limited dependence. No dependence was
defined as no methadone in the past 6 months and no IDU in the past 6 months. Limited dependence was
defined as 1-59 mg methadone daily in the past 6 months and no injecting drug use in the past 6 months. We
estimated a relative risk of injecting drug use of 9.81 (95% CI 6.76-12.86) for HIV and a relative risk of 18.53
(95% CI 15.35-21.71) for HCV (Table A4). As HBV seroconversion was not measured in the study we applied
the relative risk for HIV to HBV.
Table A4: Estimated Relative Risks
Cause
ICD
Sex
Age
RR (CI)
RR Application
HIV
Acute Hepatitis B
Liver cancer secondary
to Hepatitis B
Liver cirrhosis
secondary to Hepatitis B
Acute Hepatitis C
Liver cancer secondary
to Hepatitis C
Liver cirrhosis
secondary to Hepatitis C
A02
A11.2
B01.3.1
Both
Both
Both
15-64
15-64
15-64
9.81 (6.76-12.86)
9.81 (6.76-12.86)
Mortality & Morbidity
Mortality & Morbidity
Mortality & Morbidity
B05.1
Both
15-64
A11.3
B01.3.2
Both
Both
15-64
15-64
B05.2
Both
15-64
9.81 (6.76-12.86)
9.81 (6.76-12.86)
18.53 (15.35-21.17)
18.53 (15.35-21.17)
18.53 (15.35-21.17)
Mortality & Morbidity
Mortality & Morbidity
Mortality & Morbidity
Mortality & Morbidity
12 Figure A3: Drug dependence DALYs by region, as a proportion of all DALYS (%)
Highest
Lowest
% of total
DALYs
Glob
al
APHI
ASC
AS-E
AS-S
A-SE
Aus
Cari
bb
EurC
EurE
EurW
LAA
LAC
LA-S
LAT
Nafr
-ME
Nam
-HI
Oc
SSAC
SSAE
SSAS
SSAW
0.4
0.5
0.6
0.3
0.3
0.3
1.2
0.2
0.4
0.6
0.7
0.7
0.5
0.7
0.4
0.7
1.1
0.3
0.1
0.2
0.4
0.1
0.2
0.2
0.3
0.2
0.1
0.3
0.5
0.2
0.2
0.3
0.3
0.4
0.3
0.4
0.5
0.4
0.7
0.2
0.1
0.1
0.3
0.1
0.1
0.2
0.1
0.1
0.1
0.2
0.4
0.1
0.1
0.1
0.2
0.2
0.1
0.2
0.3
0.1
0.3
0.1
0.0
0.1
0.1
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.3
0.1
0.1
0.0
0.1
0.1
0.1
0.2
0.2
0.1
0.3
0.1
0.03
0.05
0.1
0.02
0.05
0.1
0.03
0.01
0.03
0.01
0.1
0.04
0.03
0.02
0.1
0.1
0.1
0.2
0.1
0.1
0.1
0.01
0.01
0.01
0.02
0.01
0.85
1.10
1.13
0.71
0.63
0.91
2.50
0.64
0.83
1.02
1.40
1.50
1.10
1.70
1.50
1.40
2.50
0.71
0.24
0.46
0.92
0.23
Opioids
Cannabi
s
Amphetamines
Cocaine
Other drugs
NB. AP-HI: Asia Pacific, High Income, As-C: Asia Central, AS-E: Asia East, AS-S: Asia South, A-SE: Asia Southeast, Aus: Australasia, Caribb: Caribbean, Eur-C: Europe Central, Eur-C: Europe Eastern, Eur-W:
Europe Western, LA-An: Latin America, Andean, LA-C: Latin America, Central, LA-Sth: Latin America, Southern, LA-Trop: Latin America, Tropical, Nafr-ME: North Africa/Middle East, Nam-HI: North America,
High Income, Oc: Oceania, SSA-C: Sub-Saharan Africa, Central, SSA-E: Sub-Saharan Africa, East, SSA-S: Sub-Saharan Africa Southern, SSA-W: Sub-Saharan Africa, West
13 Figure A4: Country-level DALYs per 100,000 population for cannabis, amphetamine, cocaine and opioid dependence, age-standardised, for 2010
Opioid dependence DALYs per 100,000 persons
Amphetamine dependence DALYs per 100,000 persons
Cocaine dependence DALYs per 100,000 persons
Cannabis dependence DALYs per 100,000 persons
14 Table A4 : Details of age-standardised country-level disability-adjusted life years (DALYs) per 100,000
for total illicit drug burden, 2010
Country
Global
Afghanistan
Albania
Algeria
Andorra
Angola
Antigua and Barbuda
Argentina
Armenia
Australia
Austria
Azerbaijan
Bahrain
Bangladesh
Barbados
Belarus
Belgium
Belize
Benin
Bhutan
Bolivia
Bosnia and Herzegovina
Botswana
Brazil
Brunei
Bulgaria
Burkina Faso
Burundi
Cambodia
Cameroon
Canada
Cape Verde
Central African Republic
Chad
Chile
China
Colombia
Comoros
Congo
Costa Rica
Côte d’Ivoire
Croatia
Cuba
Cyprus
Czech Republic
Democratic Republic of the Congo
mean
343.2
268.6
359.5
334.6
426.8
275.1
552.4
509.4
367
659.1
555.9
476.9
403.6
252.7
605.1
556.7
454.2
625.9
206.1
319.1
554.3
283.7
368.1
511.1
356.5
259.5
206.3
286.1
263.6
237.3
581
217.3
262.6
190
437
201.6
319.5
269
273.8
377.7
238.1
333.9
345.7
296.6
244.9
220.3
Age-standardised DALY rate (per 100,000)
Lower 95% UI
271
171.7
250
227.3
292.5
170.9
370.8
358.1
250
500.7
422.1
340.9
246.3
158.3
433.4
389.8
312.6
421.1
136.6
204.4
376.4
199.4
200.7
376.3
237
180.4
130.5
189.8
170.1
156.1
427.8
137.5
175.7
125.9
311.4
139.2
214
163.9
176.7
258.8
154.8
221
229.2
223.1
178.7
144.2
Upper 95% UI
421.3
397.5
505.2
501.5
617.6
432.9
783.8
712.6
559.8
887.7
698.8
831.1
722.6
378.4
908.8
790
642.5
927.1
298.7
489.6
889.8
388.6
881.1
669.1
518.3
356.6
309.4
420.6
393.3
342.5
760.3
321.9
396
274.5
610.2
283.5
452.1
447.5
416
535.4
361.3
461.8
507.8
382.9
322.8
325.2
15 Denmark
Djibouti
Dominica
Dominican Republic
Ecuador
Egypt
El Salvador
Equatorial Guinea
Eritrea
Estonia
Ethiopia
Federated States of Micronesia
Fiji
Finland
France
Gabon
Georgia
Germany
Ghana
Greece
Grenada
Guatemala
Guinea
Guinea-Bissau
Guyana
Haiti
Honduras
Hungary
Iceland
India
Indonesia
Iran
Iraq
Ireland
Israel
Italy
Jamaica
Japan
Jordan
Kazakhstan
Kenya
Kiribati
Kuwait
Kyrgyzstan
Laos
Latvia
Lebanon
Lesotho
Liberia
Libya
Lithuania
Luxembourg
468.3
400.5
523.5
411.6
439
332.4
332.8
838.9
297.7
752.5
254.7
486.5
358.7
413.5
333.6
476.1
322.2
397.8
221.9
378.9
593.5
414.5
207.4
195.9
397.1
449.4
378.3
267.5
455.2
292.3
287
503.5
282.4
511
381.4
485.4
370.3
364.3
307.1
389.6
307.3
700.7
427.4
537.8
430.8
479
295.9
352.4
208.8
429.6
504.1
546.6
360.4
247.1
338.8
278.6
287.2
225.8
212.4
221.1
184.4
566.2
164.7
298.7
252.4
267.1
250.3
232.2
222.2
306.9
143.5
274
386.2
263.8
137.3
133.4
274.5
312.7
240.8
187.3
322.9
205
196.2
342.7
179.1
390.8
262.5
365.5
243.4
246.9
200.5
257.2
193.8
347.3
236.6
402.2
268.4
338.3
189.7
217.3
134.7
290
376.1
415.7
624
642.9
742.5
586.4
682.2
483.2
506.1
6,015.30
488.5
991.2
384.2
848.5
526.2
528.3
434.6
1,706.80
468.5
508
332.4
498.1
809.9
640.6
299.1
284.1
586.6
618.8
612.5
377.3
646.4
391.2
417.5
816.4
419.5
645.4
549.2
637.8
540.9
526.7
464.7
671
464.7
1,531.10
953.6
761.3
652.2
729.1
452.7
558.6
309.8
675.1
708.3
708.9
16 Macedonia
Madagascar
Malawi
Malaysia
Maldives
Mali
Malta
Marshall Islands
Mauritania
Mauritius
Mexico
Moldova
Mongolia
Montenegro
Morocco
Mozambique
Myanmar
Namibia
Nepal
Netherlands
New Zealand
Nicaragua
Niger
Nigeria
North Korea
Norway
Oman
Pakistan
Palestine
Panama
Papua New Guinea
Paraguay
Peru
Philippines
Poland
Portugal
Qatar
Romania
Russia
Rwanda
Saint Lucia
Saint Vincent and the Grenadines
Samoa
São Tomé and Príncipe
Saudi Arabia
Senegal
Serbia
Seychelles
Sierra Leone
Singapore
Slovakia
Slovenia
261.8
280.7
307.2
319.2
318.6
215.8
418.5
820.4
227.7
243.6
343.8
343.5
403.8
293.4
353.1
315.9
573.3
305.4
709
320.1
639.2
332.2
194.1
206.2
233.6
630.6
402.7
312.9
286.5
413.3
300.2
331.2
437.5
251.3
357.3
320.5
654.9
280.6
664.6
269.3
442.4
524.6
810.3
211.4
423.2
210.1
348.8
617.6
208.6
338.6
264.1
284.5
175.7
174.8
188.6
209.6
215.1
137.3
316
386.2
145.3
170.4
232.3
243.6
289.1
201.9
244.6
185
421.9
195.3
457.2
232.8
452.5
217.8
119.9
135.8
166.9
476.6
235.9
225.1
186.8
276.1
195.9
220.4
289.6
160.1
247.1
216.8
258
191
429.9
168.1
300.5
353
281.3
138
243.7
134.5
246.6
314.8
136.9
223.6
201
206.8
370
444.5
451.8
487.2
475.3
327
538.2
1,896.80
333.1
357.6
509.9
478.2
559.5
421.2
507
490
764
467.2
1,136.70
417.2
895.9
488.1
292.7
297.8
316.6
862.4
930.5
423.1
410.4
621.6
454.3
491.2
661
383.9
494.3
470.5
2,584.50
404
1,123.40
388.6
627
737
1,985.30
310.7
830.8
311.7
481.1
1,670.70
301.6
500.1
356.5
369.1
17 Solomon Islands
377
260
531.1
Somalia
299.1
188.7
446.9
South Africa
680
399.6
1,278.10
South Korea
377.5
264.2
544.5
Spain
548.2
410.5
697.4
Sri Lanka
328.7
215.4
472.8
Sudan
274.2
161.9
481.2
Suriname
557.7
394.7
791
Swaziland
400.9
263
576.3
Sweden
387.3
271.5
568.6
Switzerland
383.4
265.3
543.9
Syria
294.5
188
436.7
Taiwan
259.4
186.1
367.7
Tajikistan
318.8
223
426.8
Tanzania
290.2
180.3
448.9
Thailand
311.7
225.3
445.4
The Bahamas
565.9
347.1
1,118.00
The Gambia
216.1
141.2
325.4
Timor-Leste
247.3
157.9
375.1
Togo
202.9
138
294
Tonga
981.8
296.7
2,512.80
Trinidad and Tobago
535.3
352.5
916
Tunisia
357.9
197
583.2
Turkey
383
264.5
554.4
Turkmenistan
393.4
246.4
698.2
Uganda
278.7
174.9
411.4
Ukraine
609.8
441.5
828.2
United Arab Emirates
677.2
322.6
2,393.60
United Kingdom
726.6
541.3
913.1
United States
816.2
626
1,046.20
Uruguay
445.8
309.8
636.8
Uzbekistan
587.8
427
815.3
Vanuatu
409.1
268.8
698.6
Venezuela
359.7
241.1
528.3
Vietnam
344.3
233.4
478.1
Yemen
267.2
169.8
392.6
Zambia
305.5
193.1
468.9
Zimbabwe
295.2
193.4
437.4
Note. Analyses include DALYs attributable to cannabis, cocaine, amphetamine, opioid, and other drug dependence, as well
as the following risks of illicit drug use: cannabis use as a risk factor for schizophrenia; injecting drug use as a risk factor
for HBV, HCV, and HIV; and opioid, cocaine, and amphetamine dependence as risk factors for suicide. UI: Uncertainty
interval.
18 References
1.
Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A, et al. Common values in assessing health
outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010.
Lancet. 2012; 380(9859): 2129-43.
2.
Mont D. Measuring health and disability. Lancet. 2007; 369(9573): 1658-63.
3.
Murray CJL, Lopez AD, editors. The Global Burden of Disease: a comprehensive assessment of mortality and
disability from diseases, injuries, and risk factors in 1990 and projected to 2020. USA: World Health Organization, Harvard
School of Public Health, World Bank.; 1996.
4.
U.S. National Institutes of Health National Institute on Alcohol Abuse and Alcoholism. National
Epidemiologic Survey on Alcohol and Related Conditions Wave 1 and Wave 2. National Institutes of Health,.
5.
Australian Bureau of Statistics. National Survey of Mental Health and Wellbeing of Adults 1997. Canberra,
Australia: Australian Bureau of Statistics,.
6.
Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M, et al. The Global Burden of Non-Fatal Health
Outcomes for 1,159 Sequelae of 291 Diseases and Injures 1990-2010: A Systematic Analysis. Lancet. 2012; In press.
7.
Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. Global and regional mortality from 235
causes of death for 20 age-groups in 1990 and 2010: A systematic analysis. Lancet. 2012; In press.
8.
Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. Global and regional mortality from 235
causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010.
The Lancet. 2012; 380(9859): 2095-128.
9.
Wang H, Dwyer-Lindgren L, Lofgren KT, Rajaratnam JK, Marcus JR, Levin-Rector A, et al. Age-specific and
sex-specific mortality in 187 countries, 1970-2010: a systematic analysis for the Global Burden of Disease Study 2010. The
Lancet. 2012; 380(9859): 2071-94.
10.
Degenhardt L, Hall W, Lynskey M, McGrath J, McLaren J, Calabria B, et al. Should we make burden of disease
estimates for cannabis use as a risk factor for psychosis? PLoS Medicine. 2009; 6(9): Doi:10.1371/journal.pmed.1000133.
11.
Large M, Sharma S, Compton MT, Slade T, Nielssen O. Cannabis Use and Earlier Onset of Psychosis: A
Systematic Meta-analysis. Arch Gen Psychiatry. 2011: archgenpsychiatry.2011.5.
12.
Foti DJ, Kotov R, Guey LT, Bromet EJ. Cannabis use and the course of schizophrenia: 10-year follow-up after
first hospitalization. American Journal of Psychiatry. 2010; 167(8): 987-93.
13.
Ferrari AJ, Norman R, Page A, Baxter AJ, Vos T, Whiteford HA. The burden attributable to mental and substance
used disorders as a risk factor for suicide: Findings from the Global Burden of Disease 2010 Study In progress. 2012.
14.
Ferrari AJ, Norman R, Page A, Baxter AJ, Vos T, Whiteford HA. The population attributable risk of suicide due
to mental disorders for the Global Burden of Disease 2010 Study. In progress. 2012.
15.
Van Den Berg C, Smit C, Van Brussel G, Coutinho R, Prins M. Full participation in harm reduction programmes
is associated with decreased risk for human immunodeficiency virus and hepatitis C virus: evidence from the Amsterdam
Cohort Studies among drug users. Addiction. 2007; 102(9): 1454-62.
16.
Mathers B, Degenhardt L, Phillips B, Wiessing L, Hickman M, Strathdee S, et al. Global epidemiology of
injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008; 372: 1733-45.
17.
Nelson P, Mathers B, Cowie B, Hagan H, Des Jarlais D, Horyniak D, et al. The epidemiology of viral hepatitis
among people who inject drugs: Results of global systematic reviews. Lancet. 2011; 378: 571-83.
19