Prescriptions

A systematic review of 24 studies found 56% of people who attempt to stop antidepressants experience withdrawal effects, with 46% describing those symptoms as severe. Current NICE and APA guidelines state withdrawal is "self-limiting" and resolves within one to two weeks. Seven of ten studies providing duration data directly contradict this, with symptoms lasting months to years for a significant proportion of patients.

Horowitz and Taylor used PET imaging to demonstrate a hyperbolic relationship between SSRI dose and serotonin transporter occupancy. Reducing doses in linear steps produces disproportionately large neurochemical changes at lower doses, making the final withdrawal stage the most destabilising. Withdrawal symptoms are frequently misdiagnosed as relapse, leading to reinstatement and indefinite use.

England dispensed 89 million antidepressant items in 2023/24, prescribed to approximately 8.7 million patients. This represents a 109% increase since 2010, when approximately 43 million items were dispensed. Stedman et al. (2025) confirmed this trajectory using NHS Business Services Authority data across 2010 to 2023. Antidepressants are now the second most-dispensed drug class in England, behind only cardiovascular medications.

The increase reflects both longer prescription durations and new patient volumes. This is because antidepressants are not prescribed as a twelve-week course and stopped when symptoms resolve; the average duration has extended, with millions of patients remaining on treatment for years or indefinitely. NICE guidance recommends considering discontinuation after six months for a first episode, but data shows a growing proportion of patients remain on antidepressants for over two years.

The evidence base for long-term benefit beyond one to two years is substantially weaker than for short-term treatment. A 2022 Cochrane review (Henssler et al.) found that stopping antidepressants was associated with higher relapse rates but also that discontinuation symptoms are frequently misclassified as relapse. The UK spends approximately £315 million per year on antidepressant dispensing at net ingredient cost, not including the wider NHS infrastructure cost of chronic prescription management.

Atorvastatin is the most dispensed drug in England at 65 million items in 2023/24.

Between 2004 and 2018 the NHS spent £6.3 billion on statins in primary care.

Research suggests £2.8 billion of that, approximately forty percent, could have been saved through optimal prescribing.

For primary prevention in people without existing heart disease, the Cochrane meta analysis found you would need to treat 138 people with statins for five years to prevent one death.

For the lowest risk patients now eligible under NICE's 2014 expanded threshold, the number needed to treat rises to 400.

The number needed to harm for new onset diabetes is 99 over five years. For muscle pain, approximately twenty one.

Omeprazole, a proton pump inhibitor, is the second most dispensed item at over 35 million per year.

A 2025 study across 62 English GP practices found 62% of patients on continuous PPIs had no recorded clinical indication for ongoing use.

A 2010 meta-analysis by Sattar and colleagues in The Lancet, pooling thirteen randomised controlled trials with ninety one thousand one hundred and forty participants, found that statin therapy was associated with a nine per cent increased odds of new-onset type two diabetes. The risk was greater at higher doses. The absolute risk was modest: one additional diabetes case per two hundred and fifty five patients treated for four years, but the finding has clinical significance because statins are prescribed to tens of millions of people, most for primary prevention.

The PRIMO observational study of seven thousand nine hundred and twenty four patients found myalgia in ten point five per cent of statin users, with the highest rates in atorvastatin and simvastatin users. The CTT Collaboration (2022) confirmed that benefits in primary prevention are substantially smaller than in secondary prevention, where prior cardiovascular events establish high baseline risk. A mechanism for the diabetes association involves statin suppression of CoQ10 synthesis and impairment of mitochondrial function in pancreatic beta cells, though this remains under investigation.

The standard NHS prescription charge in England is £9.90 per item in 2024, a figure that has risen consistently above inflation for four decades while the underlying medicines cost the NHS far less in many cases.

Scotland, Wales and Northern Ireland have abolished the prescription charge entirely. England's retained charge creates a geography of inequality where patients in the most deprived areas, who are also the sickest, face the highest financial barriers to medication.

Over one billion prescription items are dispensed in England annually. The charge brings in roughly £600 million per year, a rounding error against the NHS budget that functions more as a deterrent to access than a meaningful revenue stream.

The Medicines and Healthcare products Regulatory Agency receives approximately seventy five percent of its income from pharmaceutical industry fees, confirmed in a 2023 Parliamentary written answer breaking this into 50% service fees and 25% periodic fees. A 2022 BMJ investigation placed the figure at 86%. The European Medicines Agency receives 89% from industry and Australia's TGA receives 96%.

The 2005 House of Commons Health Committee warned the MHRA risked "losing sight of the need to protect and promote public health." In 2025, MPs described the funding model as "marking one's own homework" after MHRA CEO Dame June Raine characterised the agency as shifting from "watchdog to enabler." The regulator's former CEO, Ian Hudson, previously worked for GlaxoSmithKline.

Between 1998 and 2016, the total oral morphine equivalency of opioids prescribed in England increased by 127%, meaning the drugs became substantially more potent even as addiction awareness grew. By 2017/18, 12.8% of the adult population had been prescribed an opioid. Prescribing rates were nearly twice as high in the most deprived areas, with 90% of the highest-prescribing regions concentrated in the North of England.

Pregabalin prescriptions rose by approximately 900% between 2007 and 2017, with over half prescribed for off-label use. Gabapentinoid-related deaths increased from fewer than one per year before 2009 to 137 in 2015. The pattern is consistent: drugs introduced for narrow indications expand into widespread use long before the long-term consequences are understood.

For people at low cardiovascular risk, 140 patients must take statins daily for five years to prevent a single heart attack or stroke. Statins do not reduce overall mortality in this group. Yet over 7.6 million adults in England take them, with 73 million atorvastatin prescriptions dispensed in 2024/25 alone.

A meta-analysis of 91,140 participants found statins increase type 2 diabetes risk by 9%, with a number needed to harm of 255 over four years. Muscle symptoms affect 10 to 25% of users in observational studies and statins significantly deplete coenzyme Q10, essential for mitochondrial energy production. For the majority taking statins without prior heart disease, the benefit is marginal and the side-effect burden is rarely discussed.

Gøtzsche (2013) estimated that prescription drugs are the third leading cause of death after heart disease and cancer, killing approximately 197,000 Europeans and 128,000 Americans annually. These figures exclude overdoses, misuse and errors. They represent deaths caused by correctly prescribed medications taken at recommended doses in patients for whom the drugs were indicated.

This is because clinical trials used for drug approval are typically twelve to twenty four weeks long, enrolling healthy or mildly ill volunteers and powered to detect efficacy rather than rare lethal adverse effects. Post-marketing surveillance is passive and captures an estimated 1 to 10 per cent of actual adverse reactions (Hazell and Shakir 2006). Drugs that kill one in a thousand users may take years or decades to generate enough reported cases to trigger regulatory review.

The economic incentive structure ensures that drugs remain on the market even when cumulative evidence suggests net harm. Vioxx killed an estimated 60,000 Americans before withdrawal, yet Merck’s internal documents showed they knew of cardiovascular risks years before acting. The pharmaceutical industry spent $6.58 billion on direct-to-consumer advertising in the US in 2020, more than the entire budget of the FDA.

The STOMP trial (Parker et al. 2013, JAMA Internal Medicine) found that high-dose atorvastatin significantly reduced muscular strength and increased exercise-induced muscle pain in healthy adults. Observational studies report muscle symptoms in 10% to 29% of statin users, with the StatinWISE trial (Herrett et al. 2021, BMJ) finding that one-third of patients attributed muscle symptoms to their statin. Rhabdomyolysis, the most severe form of statin myopathy, occurs in approximately 1 in 10,000 patient-years.

This is because statins inhibit HMG-CoA reductase, the rate-limiting enzyme in the mevalonate pathway. This pathway produces not only cholesterol but also coenzyme Q10 (ubiquinone), essential for mitochondrial electron transport. Reduced CoQ10 impairs mitochondrial ATP production in muscle cells. Skeletal muscle is particularly vulnerable because it has high mitochondrial density and energy demand. Ghirlanda et al. (1993) showed that statin therapy reduced plasma CoQ10 levels by 40% within 30 days.

In clinical trials, muscle symptoms are frequently classified as "subjective" and excluded from efficacy analyses. The CTT Collaboration meta-analyses, which provide the evidence base for statin guidelines, used individual patient data from trials funded by the manufacturers and did not report muscle symptoms as a primary adverse outcome. When the people running the trial define what counts as a side effect, the side effect rate is whatever they want it to be.

Over 73 million proton pump inhibitor items are dispensed annually in England, making omeprazole the second most prescribed drug in the country. Long-term PPI use suppresses gastric acid, significantly depleting vitamin B12, magnesium, calcium and iron. A meta-analysis of 11 studies found PPI use increases hip fracture risk by 30% and spinal fracture risk by 56%.

The FDA issued a safety warning in 2011 about PPI-induced magnesium depletion and a 43% increased risk of hypomagnesaemia has been confirmed across meta-analyses. Magnesium deficiency is independently linked to anxiety and depression through NMDA receptor dysregulation, creating a direct pathway from one prescription to the next. Most patients are never informed their acid-blocking medication may be depleting the nutrients their nervous system requires for stability.

The Therapeutics Initiative at the University of British Columbia analysed all major statin trials for primary prevention (patients with no prior cardiovascular event). Their meta-analysis found a number needed to treat (NNT) of approximately 250 to 300 over 5 years to prevent a single death. In the JUPITER trial, the NNT was 239 over 1.9 years. In ALLHAT-LLT, pravastatin showed no mortality benefit at all versus usual care over 6 years.

This means that for every person whose death is prevented by statin therapy in primary prevention, 249 to 299 people take a daily medication for years without mortality benefit. Meanwhile, the number needed to harm (NNH) for new-onset diabetes with statin use is approximately 255 (Sattar et al. 2010, Lancet). The NNH for myopathy and muscle symptoms ranges from 10 to 50 depending on the statin and definition used. A drug that prevents one death in 300 while causing diabetes in 1 in 255 is not a clear clinical win.

Statins remain the most prescribed drug class in the UK, with over 8 million people taking them. The treatment threshold has been progressively lowered from a 30% 10-year cardiovascular risk to a 10% risk, expanding the eligible population by millions. Each threshold reduction increases the NNT because lower-risk patients receive less absolute benefit from treatment.