Thread by @SteveTransform

THREAD.
1. How do we determine the optimal drug policy model given how our policy choices impact on such a wide range of outcomes?

Simple answer: ITS COMPLICATED

So a group of experts including @olerogeberg @ProfDavidNutt & @FMeasham tried & figure out a rational new approach

2. Happy to announce the 1st paper from this long-gestating project has now been published in @ijdrugpolicy
"A new approach to formulating & appraising drug policy: A multi-criterion decision analysis applied to alcohol & cannabis regulation"
ijdp.org/article/S0955-…
open access

3. before talking about the paper - a bit of background.

Firstly, drug policy evaluations - esp for illegal drugs - has historically been poor; emphasising process measures like arrests/seizures, and in terms of outcomes, being preoccupied with prevalence of use, above all else

4. drug policy has also tended to unfold within a highly polarised and emotive political space - more often driven by tabloid hysteria and political posturing than science, pragmatism & evidence based norms

5. even when trying to be scientific/objective - it soon becomes clear that drug policy impacts on huge range of outcomes, not just levels of use & crime - but a wide array of health, social, political, personal, economic and international costs... AND BENEFITS

6. Different stakeholders will naturally prioritise different outcomes. eg
- Police > crime reduction
- Treasury > effective expenditure and tax revenue
- Public health > reduced dependence & mortality
- Parent groups > child safety
- Politicians > +ve Daily Mail coverage etc

7. ...even the most open minded policy maker will find balancing policy impacts on such a cross cutting spectrum of policy variables tricky - everyone is subject to cognitive biases reflecting there knowledge and personal priorities

8. Complicating things further, there may be trade-offs between policies and outcomes.
e.g. raise tobacco taxes/prices and tax revenues may increase and use will fall, but smuggling and tax avoidance will rise.
How to get the balance right, and who decides?

9. So...led by @Drug_Science & the Ragnar Frisch Centre for Economic Research, & funded by the Norwegian Research Council a 'Multi Criteria Decision Analysis' (MCDA - Ill explain that in a sec) was undertaken to compare the impacts of different policies on 3 different drugs

10. The thinking behind the project emerged from a critique of an earlier MCDA - the now famous Nutt et al. paper in the lancet comparing drug harms
ias.org.uk/uploads/pdf/Ne…

11. A paper by Measham and *cough* Rolles noted how the Nutt.et al analysis failed to account for the imoact of the legal/policy environment on drug harms. tdpf.org.uk/resources/publ…
Illustrating the point with the example of heroin:

12. The new MCDA tackled this shortcoming by looking at the various policy impact areas for individual drugs under 4 policy models;
- Prohibition (fullscale 'war on drugs')
- Decriminalisation of use (e.g. Portugal)
- Legal market under strict state control
- Legal 'free market'

13. A more sophisticated set of impact criteria was also devised (27 in all - divided into 7 thematic 'clusters') - and now I have to deploy the twitter nerd stick and point you to the paper if you want to see what they all are. there just isnt room here. but....

14. crucially they included benefits as well as hams amongst the impacts. Obviously there are benefits associated with both use and markets - and to ignore these is just #badscience that can only lead to distorted and sub optimal policy making

15. so the having established the 'Multi Criteria' bit of the MCDA - it was on to the 'Decision Analysis' bit. This involved what's called a decision conference - a multi discipline group of experts facilitated by an MCDA specialist.

16. For each drug (and we only looked at 3 due to tome constraints; alcohol, cannabis and heroin) the group agreed a comparative preference score for the impacts on each of the 27 criteria under each of the 4 policy scenarios (stay with me now - this is the important bit)...

17. crucially - after rating the models against each other for each criteria, the criteria were WEIGHTED against each other. This is because not all impacts are as important - we might, for example, think child safety more important than tax revenue.

18. There's more detail & discussion of the MCDA process in the paper. Please read and digest it - its genuinely fascinating. In many ways the innovative application of the MCDA process to drug policy is as important as the findings that came from process.
talking of which...

19. Here r the comparative outcomes for the 4 models (for alcohol & cannabis) summarised in graphical form. Note: unlike the famous Nutt et al harms graph these are +ve preference scores, so the larger the bar the better the model scored, on a given criteria & cumulatively

20. For both alcohol & cannabis the state control model scored highest & prohibition the lowest. Alcohol prohib scored relatively more than cannabis prohib (due to greater health/social harms of alc use - & use moderating effect of prohibition)

21. Interestingly free mkt model scored better than decrim model for cannabis - but not for alcohol (again due to the dramatically higher health and social costs of alcohol use and moderating effect of prohibiting legal supply).

22. Although not discussed in the paper - I was struck by how the MCDA data reinforces the case for effective regulation & highlights the avoidable health and social harms of under-regulation (either from prohibition or free markets).

23. The summary preference curves echo John Mark's 'paradox of prohibition' U-curve (albeit inverted) widely adapted by Transform & others. The goal of optimum regulation is the same for alcohol & cannabis, but the starting points are very different

24. In closing - it is of course important to acknowledge the limitations of the MCDA (also discussed in the paper) - not least the delphic nature of the process and how it reflects the biases of the decision conference group. But it was notable how the process had the power...

25...to change people minds and challenge instinctively held or entrenched views. It was agreed that it would usefully be applied not only to other drugs but also using different groups of experts - perhaps with more diverging views of representing different stakeholder groups

26. ..so quite aside from the important/interesting findings (explored in more depth in the paper, so dig in) I'm convinced the MCDA process has wider application in drug policy, with the power to help find common ground & pragmatic ways forward amidst the dogma and polarisation.

27. A discussion paper on the heroin MCDA is forthcoming, as are a number of related discussion pieces - follow @drugscience for updates.
And as its a way back down the thread - here's the link to the paper again (READ IT) Cheers.
ijdp.org/article/S0955-…

28. finally to note - this thread is my thoughts on the paper & not necessarily those of the other authors, project coordinators, or funders.

They are are also not responsible for the typos (gimme a break - I did this on my phone!)
:-)
ENDS/