NICE NG248: what to ask, when to refer, what to offer
- NICE guideline NG248, "Gambling-related harms: identification, assessment and management", was published by the National Institute for Health and Care Excellence on 28 January 2025, was last reviewed on that date, and has had no update published since. It covers adults aged 18 and over who gamble, plus affected others of any age.
- It contains 70 numbered recommendations across 7 sections, running from case identification through treatment to relapse and family support, and rests on 11 evidence reviews labelled A to K. The full guideline PDF runs to 56 pages.
- Free to read in full on the NICE website, including the PDF, with no registration, no account and no paywall. Reuse is a separate question: the NICE UK Open Content Licence permits free reuse in the United Kingdom only, and outside the UK any use beyond personal research and study requires written permission and is subject to a fee.
- Its subject is gambling and nothing else: gaming, problematic social-media use and smartphone use are addressed nowhere in the 56 pages. The stated population is adults aged 18 and over, plus affected others of any age, but NICE's final scope adds that the guideline may also apply to services for younger people, so it is not simply silent on under-18s who gamble.
A patient mentions debt in passing. Another has been relapsing for three months and you did not know, because nobody asked. NG248 is the document that converts both of those moments from clinical instinct into a stated procedure, and it does so without inflating the category it is describing.
The two numbers that set the frame
The guideline's Context chapter opens with figures from the Gambling Survey for Great Britain (2023), and they cut in both directions at once. 2.5% of the population aged 18 and over score 8 or more on the Problem Gambling Severity Index, the threshold NICE treats as problem gambling. A further 12% gamble at what the survey calls elevated risk of harm, scoring 1 to 7. That second band is the one most often misquoted. It is a risk gradient, not a diagnosis, and NG248 does not treat it as one.
The other figures are there to stop the first band being waved away. The Office for Health Improvement and Disparities estimates between 117 and 496 suicides a year in England associated with problem gambling or gambling disorder, and puts the excess cost to the NHS, the wider public sector and society at between £1.05 and £1.77 billion annually at 2021 to 2022 prices. An estimated 7% of the population of Great Britain, some 3.8 million adults, children and young people, are affected others who have personally felt the consequences of someone else's gambling, a figure NICE itself flags as an estimate drawn from studies using varied methods. NHS provision now stands at 15 specialist gambling treatment clinics.
Recommendation 1.1.11 is the one to read twice. Gambling and gambling-related harms can be a dominant risk factor for suicidal ideation and suicide attempts even in the absence of other risk factors. Recommendation 1.1.13 adds that the risk may be highest immediately after a gambling episode.
The treatment ladder, with doses
Section 1.5 is unusually specific for a guideline of this kind. Group CBT is offered first, to reduce gambling severity and frequency, started as soon as possible after diagnosis (1.5.13). Individual CBT is offered if the person does not want a group, if no suitable group can be formed, or if a group is judged unsuitable (1.5.14). The course is defined: usually 8 to 10 sessions for group therapy and 6 to 8 for individual therapy, delivered against evidence-based protocols, always including a relapse-prevention component (1.5.15). Motivational interviewing is a "consider" for people hesitant about starting (1.5.12), and peer support is offered as an integral part of care rather than an afterthought (1.5.11).
Pharmacology sits below psychotherapy, not beside it. Naltrexone is a "consider" only where an adequate course of psychological therapy has not achieved the desired outcome or relapses have repeated (1.5.16), started by or under the supervision of a specialist (1.5.17), with a stated regimen of 25 mg once daily for 3 days then 50 mg once daily for 4 to 6 months, and checks on renal and hepatic function and on concurrent opioid use before starting (1.5.19). NICE records that in January 2025 this was an off-label use.
One recommendation has no clinical dose at all and is arguably the sharpest in the document. Recommendation 1.3.5 states that gambling treatment and support services should be commissioned and provided without influence or involvement from the gambling industry, with no conflicts of interest between commissioners, providers and the industry.
How to use it this week
The cheapest transferable instruction comes in two strengths, and the difference between them matters. Recommendation 1.1.2 says to consider asking about gambling in the same breath as smoking, alcohol and other substances, as part of a routine health check or intake, even where there is no obvious risk factor. Recommendation 1.1.3 is the stronger of the pair and does not hedge: ask. It lists the presentations that make the question obligatory regardless of anything else in the picture: mental health presentations including self-harm and suicidal thoughts, depression, anxiety, psychosis, bipolar disorder, PTSD, personality disorder and ADHD; medicines affecting impulse control such as dopamine agonists in Parkinson's disease or aripiprazole; alcohol or substance dependence, especially cocaine; homelessness; disclosed financial concerns; safeguarding or domestic abuse concerns; a family history of gambling that harms. Recommendation 1.1.6 supplies the wording, and it is deliberately blunt: "Do you gamble?" or "Are you worried about your own or another person's gambling?".
Recommendation 1.1.14 is the practical one to have ready before you need it: blocking software for online gambling, blocking marketing messages, self-exclusion schemes for casinos, arcades and betting shops, bank-level gambling payment blocks, and agreed limits on access to money, including a family member holding finances. These are not UK-specific in principle, even where the named schemes are.
Read the organisational sections with a filter. The NHS triage route, the statutory levy and the 15 clinics describe one country's plumbing. The clinical spine, ask routinely, grade by severity including PGSI score where available, agree an aim that is typically abstinence, offer group CBT of defined length with relapse prevention, escalate to naltrexone only after psychotherapy has been given a fair trial, is portable. Open it this week if you have a patient whose finances, mood and relationship troubles have never been formally connected, and go to sections 1.1 and 1.5 first.
For the presentations NICE lists – self-harm, depression, ADHD, cocaine use, homelessness, disclosed debt – asking about gambling stops being a matter of clinical judgement and becomes a stated instruction.
The subject is gambling alone, so the guideline addresses neither gaming nor problematic social-media or smartphone use; that is an absence in the text rather than a stated exclusion, and no other NICE guideline or quality standard covers those behaviours either. Its population is adults aged 18 and over, though NICE's scope adds that it may also apply to services for younger people. Large parts of sections 1.3 and 1.4 describe UK service architecture (NHS triage, the statutory levy, 15 NHS clinics) and do not transfer as written. Reading is free everywhere, but the open reuse licence applies in the United Kingdom only: outside the UK, reproducing, translating or distributing the text beyond personal research and study needs written permission from NICE and is subject to a fee.