How do we solve a problem like medicines adherence?

It is almost a cliché to say that approximately 50% of people do not take their medicines as prescribed, a statistic that dates back to a 2002 Cochrane review^​1​. Nonetheless, this figure remains a reasonable estimate.

“It varies somewhere between about 30 to 80%, so that 50% still kind of holds as an average for long-term conditions,” says Amy Chan, a clinical pharmacist and associate professor at the University of Auckland, New Zealand, who recently co-authored a call to action to improve and enhance medicines adherence^​2​. “There is no one, not even ourselves, who will take medicines 100% of the time,” suggests Chan.

For some conditions, such as HIV and tuberculosis, adherence has been found to be higher than average, whereas “around inhaled therapies, such as for respiratory conditions, [adherence] may fall below 50%,” says Chan^​3​. Results from one objective assessment of 244 patients with COPD revealed an adherence rate of just 23%^​4​.

But with every condition — even those that are life-threatening without medicine — some people will be non-adherent. “It cuts across the board, regardless of the disease and the drug,” says Dyfrig Hughes, professor of pharmaco-economics at Bangor University, Wales, and past president of the International Society for Medication Adherence.

The result is felt by the NHS in wasted prescriptions and increased admissions, costing over £1bn a year^​5​. “People present with treatment failure,” says Hughes, “and so doctors and pharmacists don’t often recognise when this results from non-adherence.”

Instead, they may try upping a patient’s dose, or try them on a new, often more expensive, drug. “If one in two people aren’t taking their medication, then it doesn’t matter how amazing these new medicines are,” says Debi Bhattacharya, professor of behavioural medicine at the University of Leicester. “That’s money down the drain.”

The cost is also felt by the patient, in increased ill health and higher mortality. Results from one review, published in 2019, looked at adherence in people aged over 50 years and showed that good medicines adherence was associated with a 21% reduction in long-term mortality risk compared with non-adherence ^​6​.

Getting derailed

Medicines adherence has been defined by the World Health Organization as “the degree to which the person’s behaviour corresponds with the agreed recommendations from a healthcare provider”.

Over the years, as the way we conceptualise the relationship between healthcare provider and patient has changed, ‘adherence’ has been adopted over past terms, such as ‘compliance’ (see Box 1).

When considering whether a patient is adherent, there are several stages you need to take into account, including initiation, implementation and discontinuation (see Figure 1).

The Pharmaceutical Journal

At each stage, patients may not be adherent, and the reasons may differ. People may not initiate treatment because “they don’t really think their condition is serious or they don’t really believe they need the medicine”, suggests Weinman.

Often, in the initial consultation, Bhattacharya says “you can see the patient is ambivalent”. Instead of going ahead, she suggests saying something like “what I’m getting from you is that you think this probably isn’t right for you. So shall we think about other options?”

If patients do initiate treatment, they may discontinue because “maybe the medicine works quite well, the pain goes away”, says Weinman. “It’s a really quite dynamic, quite complex, picture.”

To tackle this complexity, many researchers conceptualise adherence behaviours as part of a larger behavioural model: the COM-B model, where the C stands for capability, the O for opportunity and the M for motivation, with each of these being further subdivided into two types (see Figure 2).

Alison Wright, a behaviour change researcher at King’s College London, cites weight-loss injections and antidepressants as two examples of medicines that people might not feel supported in taking. “Whether that’s on the macro level — what’s the media discourse — but also on the micro [level]. What do your partner, your parents, your friends say about people who take this sort of medicine?”

Pharmacists can be an important part of the support network that creates the social opportunity part of the COM-B model. “People who are having services, such as methadone replacement, and see the pharmacist every day — that is a fundamental input into their social support,” says Delyth James, professor of health psychology at Cardiff Metropolitan University.

When it comes to reflective motivation, “the role of identity is important,” James explains: “Do you see yourself as someone who needs to take medication?”

In qualitative work she and colleagues undertook, Bhattacharya spoke to those taking antihypertensive medicines who said, “I feel fine, and yet when I take the medicines, when I collect them, all of that is a reminder that things aren’t so good.”

Others may not believe that they have the necessary skills and struggle with confidence “in their capability to stick to a prescription, take medication as appropriate or as prescribed”, says James.

Diagnosing the problem

It is not always easy for healthcare professionals to find out what concerns patients have about taking their medications and just asking them “is not the way to do it”, Chan says. Patients may feel like they will be judged if they admit to not taking their medicines as prescribed and may want to please their healthcare provider.

“The style of delivery is really important,” agrees Wright, who says approaching patients with humility is crucial. “If someone’s been living with asthma for 20 years, they’re not a professor of respiratory medicine, but they have an awful lot of life experience of what it is to have asthma,” she points out.

Bhattacharya has been working on a screening questionnaire, called the IMAB-Q (identification of medication adherence barriers questionnaire), which has been validated in more than 600 patients prescribed medicines for cardiovascular disease prevention^​7​. The questionnaire comprises 30 statements, each of which people score between 1 (strongly disagree) and 5 (strongly agree). For example, ‘Taking my medicines is a burden to me’ and ‘I am easily distracted from taking my medicines’.

“It’s asking people which of these things are you experiencing, not what are you failing to do,” she says.

Aside from screening questionnaires, another method to identify patients who need interventions is to look at prescription and dispensing records. “We have that data to our hands, especially for pharmacists,” says Chan, who suggests that pharmacists use this to give patients a nudge, for instance: “I’ve noticed that you’ve missed a couple of your last pick-ups, has there been a change in your circumstances?”

“Sometimes it’s just a question to start the thinking,” she adds.

Relying on prescription records alone is not necessarily enough, since patients may pick up their medicine from a pharmacy but not take it; however, with increasing digitisation, several routine records could be combined. Hughes gives the example of a patient with hypertension who is being monitored. If they have periods when their blood pressure goes up or are late in collecting their medicines, “there could arguably be some way to bring that information together as a red flag”, says Hughes.

Even more sophisticated means of monitoring adherence may be necessary, particularly for medicines that are extremely important, such as those for transplant patients. Here, novel technologies can help. “There are tablet containers and blister packs available, with embedded electronic sensors that register the time when that was opened,” says Hughes.

Back on track  

Once the barriers to adherence have been identified, it is not always clear how healthcare providers should address them. There is a large body of research on adherence interventions, but a 2020 Cochrane review judged that most of the evidence is low quality^​8​.

“We definitely know that adherence interventions in general lead to small increases in adherence,” says Wright, but “people aren’t always very good at describing in their papers what they did”. This makes the interventions difficult to replicate.

More high-quality evidence is needed. “Randomised controlled trials have to be the gold standard,” says Hughes, adding that researchers “need better defined eligibility criteria for the trial, and that is something that doesn’t happen in many studies”. Without that, people who are already adherent will be included in the trial population, diluting any effect of the intervention.

In March 2025, Hughes published findings from a meta-analyses and systematic review, which aimed to identify effective interventions for improving medicines adherence. The results concluded that, although the evidence base for interventions is large, it is significantly limited by risk of bias, and it was difficult to draw definitive conclusions on which interventions are most effective^​9​.

However, the research did identify actions for implementing adherence interventions into NHS practice, all of which have evidence from multiple sources demonstrating efficacy (see Box 2).

“What we’ve done in the past is a ‘one size fits all’ and that has never worked,” says Weinman.

“Adherence interventions ought to be considered as personalised medicine,” agrees Hughes. Ideally, the barriers to a patient taking their medicines as prescribed need to be identified and then any intervention should be targeted to address them.

This is the hope for the IMAB-Q. “Each of the questions, because they’re embedded in behavioural science, they’re also linked to evidence-based behaviour change techniques,” Bhattacharya says.

For example, if somebody scores highly on a statement like ‘I worry about what other people would think if they knew I took medicines’, Bhattacharya says “you would then facilitate them to think about others whose opinion they value and those that actually approve of the behaviour”.

Wright agrees that matching the problem with the solution is crucial. “If someone’s forgetting [to take their medicines], then applying pros and cons will be useless but action planning will be really helpful,” she says.

Wright is currently working on an app to help people with asthma who are struggling with adherence to their preventor inhalers.

There is hope that artificial intelligence (AI), which can massively increase the amount of data that can be processed, might turbo-charge these kinds of apps. Still, AI is only as good as the data it gets. “I’m sure we’re going to be able to do a lot with AI but I think we’re not there yet,” says Weinman.

For now, patients are much more likely to get an automated text message reminder to take their medicines from their care provider than anything using AI; however, “providing someone with a set of prompts if they’re not motivated is just annoying”, Weinman cautions.  

What about more analogue approaches? There has been some recent controversy over pharmacies phasing out multicompartment compliance aids (MCAs), which are often requested by care agencies and can be time consuming for pharmacists to prepare.

“If [patients] don’t intend to take that medication in the first place, they’re not going to take it [if it’s in an MCA]”, says James. However, for people who are motivated “having something that helps with that psychological capacity bit, and physical capability in some ways, is really important”, she adds.

For those struggling with motivation, James says visualisation can be useful. She is working on a tool for healthcare practitioners to use with patients who do not understand the importance of taking their antihypertensives when there are no obvious symptoms for them to see in themselves. “You can show them in the brain what happens over time when you have that high blood pressure”, she says, and then what happens on medication. “You show them the difference.”

Finding opportunity

Chan says the emphasis should be on putting adherence on everybody’s radar. She cites a recent pan-European survey of different healthcare providers across 40 European countries, which showed “that there isn’t a consistent approach” in how and when healthcare practitioners tackle the problem of adherence^​10​.

“I think the call to action is, make it your problem, too,” Chan says, giving the example of a medicines use review or a vaccination: “There’s an opportunity there to talk about adherence.”

The reality is that, in day-to-day practice, these conversations are often not happening. People are stretched for time and incentives are often lacking. For example, in April 2023, NHS England removed financial incentives for primary care networks to carry out structured medication reviews. The new medicine service in England is still running, but there is no high-quality evidence to show that it helps with adherence^​11​.

Weinman, who teaches at King’s College London’s School of Pharmacy, believes good training is vital. “We have a really detailed programme there, knowing that our current students are going to be prescribers later on,” he says. Students leave with an understanding of the reasons why people do not take their medicines, and some basic communication skills to both elicit and address barriers to adherence, including a simple screener they can use with patients.

“But the system [they work in] needs to encourage them,” Weinman adds. That screener is now embedded in the electronic record system at several hospitals, including King’s — “That’s a beginning, but it’s not mandatory.” 

He also believes that healthcare practitioners need to intervene earlier. “If 50% of people are going to be non-adherent, it’s something we’ve got to expect,” he says. “We can’t be surprised six months later and then try and rally around and do something.”

Bhattacharya suggests we need to reconceptualise the way we think about medicines adherence. “I think practitioners across the board would have more buy-in if they saw it as a public health issue.” She also hopes that screening questionnaires, like the IMAB-Q, will “empower people to say the patient declined treatment”, preventing healthcare practitioners from wasting time and resources on pursuing a pathway the patient does not want to take.

We can and must do better, says Chan. Otherwise, “if we continue on this trajectory, [if] we don’t prioritise [adherence], we might be having this [same] conversation in 2075.”