You have a headache after a glass of wine too many. Your back aches from another day hunched over a keyboard. That old shoulder injury is playing up again. What do you do? There is a good chance that you will reach for the unglamorous white pills lurking in your medicine cabinet.

Paracetamol is the workhorse painkiller. GPs wrote 22.5m prescriptions for it in 2013. Around 200m packets of it are sold annually, accounting for two-thirds of the UK market for over-the-counter painkillers. It is widely viewed as cheap, safe and effective.

At around 2p per 500mg tablet, it certainly is cheap. But safe and effective? Although the potentially fatal consequences of taking a paracetamol overdose are well known, the widespread belief has been that the drug is mild and relatively safe if taken at the recommended dose. However, this is increasingly being questioned by scientists, who say that taking it over prolonged periods can have serious side-effects. That might seem a risk worth taking if it were not for recent research that suggests the drug either doesn’t work, or has only a very small effect for most people.

Paracetamol rose to prominence during the 1960s in the wake of fears that aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs) such as ibuprofen could cause gastric bleeding, ulcers and other serious side-effects. Some concerns were raised about the possibility that long-term use of paracetamol could also cause internal bleeding, but the evidence for that was mixed for many years.

However, in 2011, Professor Michael Doherty, a rheumatologist at Nottingham University, published a study looking at almost 900 patients aged 40 and older who took paracetamol, ibuprofen or a combination of both for chronic knee pain. When he compared the participants after 13 weeks, it came as no surprise that one in five on ibuprofen lost the equivalent of a unit of blood through internal bleeding. What was surprising was that so, too, had the same proportion of patients who were taking paracetamol.

“Paracetamol can actually be a very dangerous drug,” says Dr John Dickson, who retired from general practice in Northallerton, North Yorkshire, last year. “It can cause kidney and liver problems, and causes as much gastrointestinal bleeding as the NSAIDs.”

In 2013, the US Food and Drugs Administration (FDA) even issued warnings that taking paracetamol can, in some rare instances, cause potentially fatal skin conditions called Stevens-Johnson Syndrome, toxic epidermal necrolysis and acute generalised exanthematous pustulosis, which can cause the top layer of skin to become detached.

The maximum 24-hour dose of paracetamol is 4g, but as little as 5g can cause liver complications, and it can be easy to overdose accidentally by taking more than one product containing it at the same time. “I have a headache, so I’ll take some paracetamol, and I’ve got a cold so I’ll take a cold product such as Lemsip,” says Professor Andrew Moore, a leading pain researcher at Oxford University. “People don’t necessarily look at the small print.”

Last year, the FDA reduced the maximum dose of paracetamol (called acetaminophen in the US) in tablets or capsules to 325mg to reduce the risk of accidental overdoses.

In the UK, in draft guidelines issued in 2013, the National Institute for Health and Care Excellence (Nice) advised GPs that they should no longer prescribe paracetamol for osteoarthritis, suggesting it had “limited benefit” and highlighting links of higher doses to cardiovascular, gastrointestinal and renal problems. Doctors, though, were quick to criticise the new advice on the grounds that it would leave them either telling patients to simply endure their pain or lead to greater use of stronger, potentially more harmful opiate-based alternatives such as tramadol and diamorphine.

In its final recommendation last year, Nice performed a U-turn, reinstating its previous backing of paracetamol, pending the outcome of a broad review of over-the-counter painkillers by the Medicines and Healthcare Products Regulatory Agency, the results of which are due later this year.

Of course, most medicines have some side-effects, and taking them always involves balancing the possible dangers against the benefits. Perhaps the downsides of paracetamol are worth the risks? Yet a 2006 review by the respected Cochrane Collaboration found that of seven previous studies comparing paracetamol with placebos, two found no difference in pain sensations, and the others found an improvement averaging 5%, an improvement the authors described as of “questionable clinical significance”.

“For most people, it’s a placebo,” says Dickson. “It’s a bit like when they say that if you swear at your wife or your husband, you feel better. It’s the same sort of concept.”

Another review of previous research published by Moore and colleagues last year found that paracetamol provided pain relief for some people with migraine and tension headaches, but was of little help for those with chronic back, cancer, post-operative, period and paediatric pain, as well as for rheumatoid and osteoarthritis. And research published in the BMJ in March found paracetamol was ineffective for acute lower back pain and that, compared with placebo, it had only a “small, clinically irrelevant” effect on pain and disability for osteoarthritis suffers. It also highlighted evidence that those taking it regularly were almost four times more likely to have abnormal liver function test results.

Lead author Gustavo Machado, from the George Institute for Global Health at the University of Sydney, and colleagues concluded: “Our results therefore provide an argument to reconsider the endorsement of paracetamol in clinical practice guidelines for low back pain and hip or knee osteoarthritis.”

The real problem is that the old model of judging drugs on the basis of research that averages out their effects makes little sense when these can vary dramatically between individuals. “What we’re recognising now is that with paracetamol, as with all analgesics, there are some people for whom it can provide good pain relief and others in whom it has no effect at all,” says Moore. In such cases, perhaps it makes sense for patients to take a greater role in managing their own treatment, working with medical professionals to find out what works for them.

Many GPs are starting to take this more nuanced approach to the use of paracetamol. Two years ago when Nice was considering withdrawing its backing of the drug as the first choice treatment for chronic osteoarthritic pain, the Royal College of General Practitioners was among those who complained the loudest. Now, however, Dr Martin Johnson of the RCGP says the status quo that drives GPs to prescribe prolonged use of paracetamol for millions of these patients no longer makes sense.

“If you look at asthma or diabetes, these are really well self-managed conditions because people are empowered, but we’re not used to the concept of self-management of pain. In fact, patients shouldn’t be using paracetamol habitually. Rather, they should take them when they have pain and when when they’re going to do something that normally provokes pain, such a going for a long walk. They should also consider other ways to manage pain such as hot baths and stretching exercises.”

Dickson agrees that a difficult shift in the mindset of doctors and patients on how to use paracetamol and other painkillers is overdue. “Doctors have traditionally said, ‘No, you shouldn’t have pain, and we’ll give you something to stop it,’” he says. “What we have to get across to people is that pain itself isn’t doing any harm, it’s not something you can cure and sometimes it makes sense not to take anything. We have to learn to manage pain more proportionately.

“It’s not going to be easy because it’s a cultural problem, but my view is that within five years we will no longer be prescribing paracetamol for chronic pain relief.”