The panel — three composite experts, twenty years each, defending the thesis

His position

Addiction is real, and the brain-disease frame — whatever its limitations — is the only frame that has unlocked treatment at the institutional scale required to reach the population that actually needs it. Twenty years of opioid substitution work in inner-city addictions services have given Marcus a particular view of the willpower argument: it is a view that does not survive contact with the patient who has been clean for eleven months, has built a life back, has a flat and a job and a child to collect from school, and then walks past a familiar street corner and is using by Tuesday afternoon. That patient did not lack willpower. He had eleven months of it. Something else happened, and that something else is what the field is trying to describe when it uses the word addiction.

What he defends on the site

The Pathway page in particular — the mesolimbic mechanism, the wanting/liking dissociation, the allostatic-load picture, the D2 imaging data. Also the opioid agonist therapy evidence on the Treatments page: the Sordo et al. BMJ 2017 mortality data is, in Marcus's view, the single most important finding in addiction medicine in his career, and the operational failure of US prisons to provide methadone or buprenorphine is, in his view, "policy-mediated mass mortality dressed as moral seriousness".

Where he pushes back on Eleanor

Marcus has limited patience for the Snowdon arithmetic on the Costs page. He concedes the lifetime-cost finding is real and that the IEA papers are technically defensible as Treasury accounting. He will not accept "they die earlier so it works out" as a public-health frame. Twenty years of writing death certificates has produced a clinical view of QALYs that is harder to monetise than the IEA model allows. Eleanor finds him slightly sentimental on this; he finds her slightly cold. They are both right.

Where he pushes back on Inês

Marcus is broadly sympathetic to harm reduction and has helped run a needle exchange. He is more cautious than Inês is about full decriminalisation in the absence of treatment capacity. He has watched the Portuguese SICAD-era cuts in real time through professional contacts and his summary is sharper than the standard advocacy line: "the model only worked when it was funded. Decriminalisation without funded treatment is just legalisation by the back door, and we know what that does to the most vulnerable users". Inês broadly agrees but holds out for the carrot-and-stick architecture as a defendable design even under stress.

His signature claim

The argument that addiction isn't real lasts until you have looked into the eyes of someone in their fourth week of involuntary withdrawal, asking you, with full insight and full distress, why their body is doing this to them. The question is not rhetorical. They want a mechanism. They are owed one. The mechanism is the disease. The word is awkward and the politics are awkward and the frame is not perfect, but the alternative is to say to the patient, in front of you, that they are doing this to themselves on purpose. That is not a position any competent clinician will hold for long.

Her position

The cost-of-illness numbers in the addiction-policy literature are sloppier than the field admits. Eleanor's twenty years inside the building have given her the practitioner's instinct: cost figures are political artefacts as much as analytical ones, and the direction of inflation is rarely random. She holds that Snowdon is partially right — the gross-versus-net distinction is structural, the van Baal 2008 finding is robust, the lifetime-cost accounting is what it is — and that the public-health response of refusing to engage with this on grounds of moral discomfort is itself a kind of bad faith. She also holds that the Snowdon corollary — that early death is a Treasury win — only works if you price QALYs at nil, which no honest analyst does. The defensible policy frame is welfare loss, not fiscal crisis. The current rhetoric on both sides is undisciplined.

What she defends on the site

The entire Costs page is hers. The "where Snowdon is right / where Snowdon deflects" structure is the structure she uses internally when colleagues ask her about the IEA papers. She is also the load-bearing author of the minimum unit pricing and fat-shaming campaigns sections on the Policy page, both of which she has watched designed and evaluated in real time.

What she most wants people to take from her sections

Three things. First: "education" is the lowest-yield lever in addiction policy and the most over-funded. The DARE evidence base is settled. We continue to fund variants of it because the political economy of "stop telling children about drugs" is impossible, not because the evidence supports doing more of it. Second: sin taxes are not health interventions, they are regressive transfers, and pretending otherwise is intellectually dishonest. The political case for them can be made on their actual mechanism — they raise revenue from a population that produces a fiscal externality — without dressing them up as compassionate public health. Third: cost-of-illness inflation is now structural in this literature. Every recalculation has produced a larger number. Treat each new figure with the scepticism Snowdon brings, regardless of which side of the argument the figure favours.

Where she pushes back on Marcus

Eleanor thinks Marcus over-uses the brain-disease frame as if it settles the policy argument. Her view: even granting the biology in full, the policy question is what we collectively choose to fund, how we choose to fund it, and on what evidence base. The brain-disease frame can produce both excellent policy (parity insurance, reduced criminalisation, expanded substitution therapy) and bad policy (over-medicalisation of normal behavioural variation, unbounded growth in the clinical perimeter, vendor capture of public-health institutions). The frame is not self-executing. Twenty years in the Treasury have left her unwilling to let any clinical framework do the policy thinking on its behalf.

Where she pushes back on Inês

Eleanor is in principle sympathetic to harm reduction and has co-authored two cost-effectiveness papers on supervised consumption. Her practitioner's caution is about scale-up. The published harm-reduction evaluations are mostly small, embedded, well-funded pilots with motivated local stakeholders. Whether the same effect sizes survive national-scale implementation under austerity is the question Eleanor keeps asking and Inês keeps half-answering. The Portuguese 2012-onward arc is, in Eleanor's view, the empirical answer to her own question.

Her signature claim

The single largest source of bad addiction policy of the last forty years has been the assumption that an information deficit is the binding constraint. It is not. People know the calorie content of a Big Mac. People know the carcinogen profile of a cigarette. People know what fentanyl does. The question every addiction policy needs to answer is what the binding constraint actually is, in the specific population it is targeting, and that question is empirical and population-specific. We have spent fifty years writing curricula instead of asking it.

Her position

Both the disease frame and the moral frame fail without operational harm reduction in place. Abstinence is the exception, not the rule. Inês's twenty years in Portuguese drug services — from the optimistic IDT era through the 2012 austerity cuts to the chronic underfunding of SICAD and its successor agency — have given her a particular view that neither Marcus nor Eleanor quite shares. The mechanism is real and the policy arithmetic is real, but at the level of the user in front of you on a Tuesday night in Bairro Alto, what is needed is the operational thing: the needle, the dose, the bed, the food, the door back into the medical system, the safe place to come back to next week. The theory matters only if it produces the operational thing. The Portuguese model did, briefly, at scale, when it was funded. When the funding was cut, the operational thing stopped happening, and overdose deaths rose. Twenty-three years on, the evidence is clear: the policy is only as good as its operational layer, and the operational layer requires sustained money. The carrot has to be funded. The stick on its own is just the war on drugs in a different uniform.

What she defends on the site

The Managed Alcohol Programs and the Portuguese decriminalisation arc on the Treatments page, in particular the candour about the 2012 backsliding. She is more honest than the standard harm-reduction advocacy literature about what happened in Portugal after the cuts. The model did not "fail" in the loose Atlantic-magazine sense — Portugal still has a fraction of the US overdose rate — but it lost capacity, the consequences are measurable, and the field needs to be able to talk about that without ceding the rhetorical ground to the prohibitionists.

What she most wants people to take from her sections

Two things. First: the policy choice is not between "addiction is real" and "addiction is willpower". It is between "we will pay for the operational layer that makes either frame deliverable" and "we will not". A well-funded prohibitionist regime can keep some users safer than an unfunded harm-reduction one. The frame matters less than the funding. Second: she is sceptical of the Anglo-American "Portugal model" advocacy that treats the 2001 reform as a magic bullet. The reform worked because the country built an entire treatment infrastructure around it, paid people, trained them, ran the CDTs as competent panels rather than rubber stamps. Anyone proposing to copy the model elsewhere without copying the infrastructure is selling a slogan.

Where she pushes back on Marcus

Inês thinks the brain-disease frame is true at the level of mechanism and partially counterproductive at the level of patient identity. Marc Lewis's argument that the disease label produces learned helplessness is not, in her experience, a marginal effect. Her users have repeatedly told her, in three languages, that the disease label can become "the thing I am", and that becoming the thing makes recovery harder. She would soften the frame at the patient interface even where she accepts it as scientific description. Marcus thinks this is a tactical mistake that undercuts the political case for treatment funding. They have had this argument before.

Where she pushes back on Eleanor

Inês finds Eleanor's economist register clarifying but slightly bloodless. The published evaluations Eleanor is suspicious of are usually small because the operational reality is that harm-reduction services run on shoestring budgets and produce evaluations to match. The absence of large RCTs is itself a policy artefact — nobody wants to fund the trial that might show the model works at scale, because the political cost of having to fund the scaled-up programme is too high. Eleanor accepts this but says it does not change the analytical job.

Her signature claim

I have buried colleagues' patients. I have also kept colleagues' patients alive for years by doing operationally unglamorous things — picking up the phone, getting them a bed, getting them their methadone, getting them a meal, knowing their name. The frame argument is a debate I am happy to have in the evening. The operational layer is the thing that produces the body count, one way or the other. We have very good evidence on what the operational layer should look like. We have very bad commitment to funding it. That is the actual problem.