If I prescribe sleeping tablets to 25 people, one of them will die. Could you live with that?

As a GP, I know how difficult it is to say no to a patient. But the latest research from Public Health England on the ‘number needed to kill’ shows why doctors need to find their backbones

The latest research on sleeping tablets has revealed that people who take them them are 3.5 times more likely to die than those who do not. That means that for every 25 people taking them, there is a death every seven years.

As a GP, I can put it more simply: if I prescribe sleeping tablets to 25 people, one of them will die. Among doctors this is known as the "number needed to kill".

The harm doctors cause by prescribing addictive medication is an issue I have been following since I first qualified. On my first day in the job, I was sitting in with the pharmacist shuffling through repeat prescriptions signed by my new colleagues. On these printed scripts I saw combinations of painkillers and high dose sleeping tablets that made me wince. The pharmacist shrugged. "Happens all the time. I can’t do anything about it, I have to issue it."

Since then I have regularly come across patients on these kinds of medicines. The hardest to cope with is frail older people, prescribed sleeping tablets by some doctor well over three weeks ago – often more like three years ago – and it is left to their current GP to untangle the mess created. Patients keep asking for something to help them sleep and doctors keep prescribing – and now we have 2.4 million active prescriptions for sleeping tablets, according to the just-published Public Health England (PHE) review of prescription drug dependence.

We were warned by America’s opioid epidemic but its not just opioids that are addictive, but sleeping tablets too, as well as two new painkillers called gabapentin and pregabalin.

Doctors are well aware of the addictive nature of these tablets – but they do not necessarily tell their patients. And I think, as a group, doctors have not yet accepted responsibility for the harm we cause when we prescribe addictive pills.

PHE recommends that the guidance on this matter should be "enhanced" – but that won't help when doctors are simply not following the existing advice. It’s not exactly a new discovery that sleeping tablets are addictive. And as early as 2004, NICE guidelines specifically stated: "Do not offer a benzodiazepine sleeping tablet for the treatment of general anxiety disorder in primary or secondary care except as a short-term measure during crisis. Follow the advice in the British National Formulary (BNF) on the use of benzodiazepines in this context."

If you turn to the BNF (available free on-line and used daily by every doctor in the UK) it reiterates that these drugs are only for "short term relief (two to four weeks only) for severe disabling anxiety". For sleep, they should not be prescribed indiscriminately but only for short courses in the "acutely distressed", that "tolerance develops within three to 14 days", and that they should always be avoided in the elderly due to the risk of sedation and falls.

You don’t need to be a doctor to understand any of that advice, and yet some doctors are still not following it.

If someone asks me for sleeping tablets, I immediately explain that they are addictive. That they do not make people feel fresher the next day and will often cause a drug hangover in the morning. Patients then usually explain to me that they are not, in fact, addicted to them, but they just need them to help them sleep. I listen. And then I say no.

I have had this conversation with patients from every part of our society; from intimidating bankers who want their own way, to brow-beaten ex-offenders recently released from prison. Saying "No" is a very uncomfortable experience for a doctor; maybe some doctors just can’t quite bring themselves to do it.

But they should. They are not drug dealers. And for every 25 patients they prescribe sleeping tablets to, one patient will die.

The guidance is quite clear; it is doctors who are not being honest with to their patients.

For me, the most important recommendation in the PHE review is to improve "information for the patients". If we can’t trust doctors to inform their patients each and every time they prescribe an addictive drug, let’s leapfrog the doctors and give the information straight to the patients themselves.

PHE has already started this with one practical measure – labelling drug boxes. Instead of the lukewarm remark "may be habit forming" used previously, new prescription boxes will contain a prominent yellow direction, in the style of a cigarette carton, saying: "WARNING: THESE DRUGS CAN CAUSE DEPENDENCE, ADDICTION AND OVERDOSE." You can't miss it. The government is telling you, just in case your doctor forgot.

Its damning, its embarrassing for the profession, but it does avoid the harm that some doctors are causing. And I, for one, welcome that.