Transcript Psychoactive drugs: modelling their harm and

Psychoactive drugs: modelling their
harm and policies for their control
Professor Lawrence Phillips
London School of Economics & Political Science
and
Facilitations Ltd
Operational Research Society
Criminal Justice Special Interest Group
18 November 2015
2
March 2007
Nutt, D., King, L. A., Saulsbury, W., & Blakemore, C. (2007). Development of a rational
scale to assess the harm of drugs of potential misuse. [Journal]. The Lancet, 369, 10471053.
3
May 2007
▫ Blakemore, Nutt and Phillips meet for first time
▫ Phillips explains MCDA, multi-criteria
decision analysis, as a possible
approach to modelling drug harms
▫ We draft research project proposal
March & June 2009
▫ Medical Research Council and
Home Office co-sponsor research project
▫ Advisory Council on the Misuse of Drugs,
(ACMD), David Nutt as Chair, meets to develop
an MCDA model and to test its potential for
evaluating drug harms
4
July 2009
5
October 2009
6
January 2010
▫ Nutt sets up the Independent Scientific
Committee on Drugs ISCD
June 2010
▫ ISCD builds on the ACMD work by completing
the evaluation of harms; this is reported here
July 2010
▫ ACMD publishes the MCDA framework
developed in 2009
▫ http://www.homeoffice.gov.uk/publications/drugs/acmd1/ACMDmulti-criteria-report
7
Nutt, D. J., King, L. A., Phillips, L. D., & on behalf of the Independent Scientific
Committee on Drugs. (2010). Drug harms in the UK: a multicriteria decision
analysis. The Lancet, 376(1558-65).
8
Decision Conference
• A one-to-three-day workshop
• To resolve important issues of concern
• Attended by key players who represent the
diversity of perspectives on the issues
• Facilitated by an impartial specialist in group
processes & decision analysis
• Using a requisite (just-good-enough)
model created on-the-spot to help
provide structure to thinking
9
Decision conference + MCDA
• A methodology for a
group of key players to
appraise options on
multiple criteria, and
combine them into one
overall appraisal
• MCDA converts all
input evaluations of
decision outcomes into
the common currency of
value added
10
MCDA provides a way
to compare apples and
oranges, provided
there is a context that
establishes added
value.
For the ISCD, it was
negative value:
physical, psychological
and social harm
11
PrOACT-URL framework
• Problem
• Objectives
• Alternatives
• Consequences
• Trade-offs
• Uncertainty
• Risk attitude
• Linked decisions
12
Methods
Study design
▫ 16 harm criteria
developed by
ACMD 
▫ 20 drugs
▫ ISCD members
plus 2 external
experts
▫ Meeting facilitated
as a decision
conference
13
The 20 drugs
Heroin
Crack
Cocaine
Alcohol
Tobacco
Amphetamine
Mephedrone
Buprenorphine
Benzodiazepines
Cannabis
Anabolic
Steroids
Ecstasy
Ketamine
LSD
Mushrooms
Methylamphetamine
Khat
Butane
Methadone
GHB
14
Scoring the drugs
• The most harmful drug on
each criterion was scored
at 100.
• All other drugs were scored
relative to that drug.
• E.g., a drug considered half
as harmful was given a
score of 50.
• This creates a unique ratio
scale for each criterion.
100 ┬ Most harm
┤
80 ┤
┤
60 ┤
┤
40 ┤
┤
20 ┤
┤
0 ┴ No harm
15
15
16
Weighting the criteria
• Some criteria represent more
harm than others.
• Swing-weights equate the
units of harm on all the
criteria: the swing in harm
from the ‘no harm’ point
to the ‘most harmful’ drug.
• The group considered this
question to compare the levels
of ‘most harm’ on the criteria:
▫ “How big is the difference in
harm and how much do you
care about that difference?”
100 ┬ Most harm
┤
80 ┤
┤
60 ┤
┤
40 ┤
┤
20 ┤
┤
0 ┴ No harm
17
ISCD UK results 2010
18
European replication May 2013
19
UK 2010 vs. Europe 2013
80
70
60
r = 0.993
50
UK
Supports the
reliability of
judgements
40
30
20
10
0
0
10
20
30
40
Europe
50
60
70
80
20
Correlations of ISCD scores with...
...van Amsterdam population
...van Amsterdam individual
3
VA individual results
VA population results
3
2
linear r = 0.84
1
exponential r = 0.88
0
2
linear r = 0.83
1
exponential r = 0.91
0
0
20
40
60
ISCD results
80
0
20
40
60
ISCD results
Reference: van Amsterdam, J. G. C., Opperhuizen, A., Koeter, M., & van den Brink, W. (2010). Ranking the harm of
alcohol, tobacco and illicit drugs for the individual and the population. European Addiction Research, 16, 202-207.
80
211
Correlation of UK Drugs Act classification
with ISCD results
UK Drugs Act classification
1A
2B
linear r = 0.04
3C
without the unclassified drugs:
linear r = 0.38
4U
5
0
21
20
40
ISCD results
60
80
22
Alcohol

Harm
to
Others
Harm to Users
23
Conclusions
• High correlations among the UK, EU and Dutch
studies establish the validity of the MCDA
approach to modelling the harm of drugs.
• The improved scoring and weighting in MCDA
increases the differentiation between most and
least harmful drugs.
• The present UK drug classification system is not
simply based on considerations of harm.
• Targeting alcohol harms is a valid and necessary
public health strategy.
24
DrugScience
18-19 July 2013
255
MCDA for 12 nicotine products
Manufactured and hand-rolled cigarettes in which the tobacco is
wrapped in paper.
Cigars
Smoked cigars: rolled tobacco leaf.
Used like a cigarette but wrapped in tobacco leaf (a product
Small cigars
largely limited to the USA but widely used there).
Pipes
A tube with a small bowl at one end for smoking tobacco.
Water pipe
Where tobacco smoke is bubbled through water.
Smokeless
Non-snus (and other) smokeless refined tobacco products used
refined
orally, including leaf chewing tobacco and snuff.
Smokeless
Non-snus (and other) smokeless unrefined tobacco products used
unrefined
orally, including leaf chewing tobacco and snuff.
Snus
A low nitrosamine and non-fermented smokeless tobacco product.
ENDS
Electronic Nicotine Delivery System products, e.g. e-cigarettes.
Oral products Orally-available nicotine products.
Patch
Dermal nicotine delivery products.
Nasal sprays Nasal nicotine delivery products.
Cigarettes
26
ISCD results 17-18 July 2013
100
90
80
70
60
50
40
30
20
10
0
TO USERS 67
TO OTHERS 33
27
DrugScience
10 September 2015
28
Context
• DrugScience collaboration
▫ Frisch Centre and SIRUS - the Norwegian Institute for Alcohol
and Drug Research (Ole Rogeberg is the lead researcher)
▫ Funded by the Norwegian Research Council.
• Purpose
▫ Develop an analytic framework for describing, measuring,
assessing and discussing drug policy
• Decision conference
▫ 10-11 September in London
▫ Develop MCDA model for comparative analysis of different
policies
29
Four policy options to test MCDA model
• Options constructed from combinations of policy features
▫
▫
▫
▫
▫
▫
Production
Sale/distribution
Purchase
Purchase volume (for legal users)
Possession
Use
• Options
▫
▫
▫
▫
Absolute Prohibition
State Control
Decriminalisation
Free Market
30
The drug policy MCDA model
• Seven impacts
▫
▫
▫
▫
▫
▫
▫
Health
Social
Political
Public
Crime
Economic
Costs
• 27 Policy criteria (e.g. Health)
▫
▫
▫
▫
▫
Harm to user
Harm to others
More harmful substance
Encourages treatment
Product quality
31
Results for alcohol
• State control most
preferred
• Key impacts
▫ Health (harm to
user)
▫ Public (protects
children, young &
vulnerable)
▫ Crime (like Free
Mkt)
• Free market least
preferred
32
State control vs.
Free market
State control
better
Free market
better
33
State control vs.
Decriminalisation
State control
better
Decriminalisation
better
34
State control vs.
Absolute
prohibition
State control
better
Prohibition
better
35
Next steps
1. Decision Conference, 20-21 January 2016
2. Test on other substances
3. Add criteria that are substance specific
(ketamine)
4. Delete criteria that don’t discriminate very well
5. Refine policy options
6. Test with other groups of experts
7. Use model to compare policies of other
countries
8. …..
36
A guide to further reading
Harvard University Press,
1992.
Shows how to articulate
values and make wise
decisions.
Dodgson, J., Spackman, M.,
Pearman, A., & Phillips, L.
(2000). Multi-Criteria Analysis:
A Manual. London: Department
of the Environment, Transport
and the Regions, republished
2009 by the Department for
Communities and Local
Government. Download free at
http://eprints.lse.ac.uk/12761.
MCDA is described in Chapter 6.
CRC Press , 2014.
Overview of B-R
assessment across the
R&D spectrum.
Chapter 5 describes
methods & applications,
including MCDA.