Article Text

Download PDFPDF

Adherence in childhood asthma: the elephant in the room
  1. Robert W Morton1,
  2. Mark L Everard2,
  3. Heather E Elphick3
  1. 1University of Sheffield, Academic Unit of Child Health, Sheffield Children's Hospital, Sheffield, South Yorkshire, UK
  2. 2School of Paediatrics & Child Health, University of Western Australia, Perth, Australia
  3. 3Department of Respiratory Medicine, Sheffield Children's Hospital, Sheffield, UK
  1. Correspondence to Dr Robert W Morton, Clinical Research Fellow, University of Sheffield, Academic Unit of Child Health, Sheffield Children's Hospital, Sheffield, South Yorkshire S10 2TH, UK; r.w.morton{at}


Adherence to inhaled steroids is suboptimal in many children with asthma and can lead to poor disease control. Many previous studies in paediatric populations have used subjective and inaccurate adherence measurements, reducing their validity. Adherence studies now often use objective electronic monitoring, which can give us an accurate indication of the extent of non-adherence in children with asthma. A review of the studies using electronic adherence monitoring shows that half of them report mean adherence rates of 50% or below, and the majority report rates below 75%. Reasons for non-adherence are both intentional and non-intentional, incorporating illness perceptions, medication beliefs and practical adherence barriers. Interventions to improve adherence in the paediatric population have had limited success, with the most effective containing both educational and behavioural aspects.

  • Child Psychology
  • Respiratory

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Imagine if we read about a new drug on the market to treat asthma. The drug in question has been tested in many clinical trials across the globe over the last 50 years, and the results are unequivocal. The Global Initiative for Asthma1 has reviewed all the data and tells us that this drug is proven to reduce symptoms, improve quality of life, improve lung function, reduce airway hyper-responsiveness, control airway inflammation, reduce frequency of exacerbations and reduce mortality. We would be naturally very excited about this drug and its potential to transform the lives of our asthmatic patients. It is perhaps with some disappointment that we then discover this is not a revolutionary new medication, but the simple inhaled steroid, a drug we have prescribed all our careers.

As paediatricians we know the benefits of inhaled steroids, but still see many of our patients on this medication suffer frequent symptoms and asthma exacerbations. We know the medication is usually effective if taken properly and regularly, so it leaves us with the question: Is this patient taking the medication as prescribed? We ask our patient and their parent in clinic if they take their inhalers and they say ‘yes’. The published literature now tells us that the majority of patients do not take their inhalers as prescribed,2 even if they report they do, but we are often still reluctant to push the issue beyond an informal enquiry. In paediatric asthma clinics, the issue of adherence soon becomes ‘the elephant in the room’.

In order to be able to address the issue of non-adherence, we need to know the extent of the problem in our paediatric population. What measures are available to us to accurately measure adherence, and are they reliable? When objective methods have been used, what are the published levels of non-adherence in children with asthma? If children are not taking their inhalers, what are the clinical consequences? Finally, and perhaps most importantly, why don't children take their inhalers, and what can we do to improve adherence? This review aims to address all of these issues to give better insight into this significant issue.

Definition of adherence

In 2003, the World Health Organisation defined adherence as “the extent to which a person's behaviour—taking medication, following a diet, and/or executing lifestyle changes, corresponds with agreed recommendations from a health care provider”.3 An adherence rate is usually quoted as the percentage of prescribed doses actually taken by the patient.4

Adherence has superseded the term ‘compliance’, which is the amount of medication taken as prescribed, implying a unilateral and dictatorial process.5 The agreement on a medication regimen between prescriber and patient is essential as it incorporates the needs and views of both the health professional and the patient, and is therefore more likely to be followed.

A third term, ‘concordance’, is increasingly used in the UK. Concordance implies an ongoing partnership between prescriber and patient. This incorporates all aspects of medication, including the initial prescription and support to facilitate taking the medication.5 Any medication regimen should be agreed upon by both parties, and regular assessments and adjustments made in order to achieve optimal success. While this term goes furthest in acknowledging the importance of full patient involvement, it is yet to be universally recognised as there are currently no objective and comparable measurements of concordance available.5

Methods of adherence monitoring

There are many ways of assessing adherence, each with varying degrees of objectivity and validity. Often the only measure of adherence we have in clinic is the direct questioning of children and their parents. Answers such as ‘he remembers it most days’ or ‘he sometimes forgets the morning dose before school’ give us some insight but cannot give an accurate picture. As clinicians, we might think we can identify all the patients who do not accurately report their adherence, but we cannot. In a study of 51 asthmatic children, clinicians only correctly predicted 55% of the patients who were inaccurate in their adherence reporting, when levels were electronically monitored.6

People want to be looked upon favourably and do not want to admit to non-adherence. This leads to over-reporting and is known as the social desirability bias. Self-report questionnaires aim to reduce this bias by making the process more confidential and less personal by the use of indirect questioning.7 Questionnaires are subjective however and have been shown in numerous studies to overestimate adherence compared with more objective methods.2 ,6 ,8–10 One American study of inner-city children showed that parents’ reported adherence was 85%, compared with under 25% when measured more objectively.11 The overestimation in questionnaires is partly still due to the social desirability bias, but also due to inaccurate recall, and generalising behaviour over the time period rather than specific events.12

Daily diaries are still subjective, but aim to avoid recall inaccuracies by providing day-to-day recording. However, diaries have also been shown to overestimate adherence rates compared with more objective measures.13–16 In one study, median parent reported adherence was as high as 95%, compared with just 58% when measured electronically.14 An additional bias is the patients who fill out diary cards regularly are also the patients who are more likely to be adherent to their medication.

Prescription filling is more objective and is commonly used in clinical practice when attempting to identify non-adherence. Like the other two methods, prescription filling has also been shown to overestimate adherence.8 There is no guarantee that the dispensed medication is taken, and one study reported adherence levels of 70%, compared with 50% when measured electronically.8 Canister weight is often cited as a reliable objective measurement of adherence and is a cheap method of analysis in research studies.8 Prescription refill, canister weight and dose counters are all subject to social desirability bias via the phenomenon of ‘dumping’, where patients activate their inhaler many times just before clinic,17 leading to overestimation of adherence.18 When measured electronically, one adult study reported that 18% of the patients showed some evidence of dose ‘dumping’, with one patient using the inhaler over 200 times on the morning of the clinic.10

Electronic monitoring devices (EMDs) record the exact time and date that an inhaler is used, and are the most objective method of adherence monitoring (figure 1). They are not yet universally acknowledged as the gold standard of adherence monitoring12 ,19 due to some early studies reporting high rates of device malfunction.9 ,18 ,20 ,21 With technological advances however, EMDs are much more reliable and have shown a high degree of accuracy in bench22 ,23 and clinical studies.24 EMDs are more expensive than other methods of monitoring, which may limit their use in clinical practice, but adherence studies are increasingly using EMDs due to their unrivalled objective validity.

Figure 1

Nexus 6 ‘Smartinhaler’.

True adherence to inhaled devices involves both taking the right quantity of medication and in the correct way. With all methods of adherence monitoring, there is no guarantee that when the medication is taken it is inhaled effectively. Children are often unable to take inhalers competently even if they demonstrate good technique in clinic. This may be due to forgetting instructions or knowing the correct method of use but choosing a different technique due to ease or social constraints.25

What are the adherence levels to inhaled steroids in children with asthma when recorded electronically?

When many studies have recorded adherence levels to preventer inhalers in children, the methods used are often subjective and therefore potentially inaccurate. Recent studies that have recorded adherence electronically give us an objective indication of adherence levels in the paediatric population. Table 1 pools the adherence data from these trials. Trials were found by searching the PubMed database for the period January 1980 to October 2013 with the terms adherence/compliance and asthma and electronic and monitoring. Other studies were identified from reference lists from these studies. Studies were included if they were monitoring adherence only in children (aged <18). Where the trials involve an adherence intervention, only the control arm rates or rates at baseline are quoted. Studies were rejected if they looked at rates in populations that were already shown to be non-adherent. The mean rate of adherence is the overall rate of adherence for the whole duration of the respective studies. If multiple adherence rates were recorded over time, the mean of these rates has been calculated.

Table 1

Mean adherence rates to inhaled corticosteroids when measured electronically

These studies show that overall adherence in the paediatric population is low when recorded objectively. Half of the studies record adherence rates of less than 50%. All of the studies except one report adherence rates below 75%. Adherence rates are lower in older children21 ,35 and in children from a poor socioeconomic background.27 ,29 ,31 ,33 ,35 The populations where adherence rates are highest are developed western countries such as the UK, the Netherlands, Australia and the USA, although these rates are still below 75%. The studies with more adherent populations also only included younger children.6 ,15 ,24 ,32

These adherence rates in research studies are likely to be higher than ‘real life’ because children and parents know the data are being recorded. This behaviour is known as the Hawthorne effect.36 In all the studies where adherence rates were recorded over a longer period, adherence rates decline over time, and the later lower rates are likely to represent the ‘real’ adherence as the Hawthorne effect wanes.8 ,24 ,26 ,28 ,30 ,34

What are the consequences of non-adherence?

As the data show, the majority of children do not take their inhalers, and in many populations they only take their inhalers half of the time. Inhaled steroids have been shown to improve lung function, decrease symptoms and reduce the need for β agonists. It can take up to 6 weeks for inhaled steroids to have a full effect.37 ,38 Regular inhaled steroids are more effective than intermittent treatment during exacerbations at decreasing lung inflammation, and improving lung function and asthma control.40 If children are not taking their inhalers regularly, theoretically they are therefore more likely to have poor asthma control, poor lung function, suffer more symptoms and use β agonists more regularly. Potential exceptions to this are children where the diagnosis of asthma is uncertain or in the very mild asthmatics.30 ,40

Population studies in adults using prescription refill data have shown that in order to significantly reduce the chance of an asthma exacerbation, adherence to inhaled steroids needs to be in excess of 75%.41

Studies have shown that patients with low adherence rates are indeed more likely to have poorly controlled asthma.30 Children with poor adherence are more likely to require courses of oral steroids14 ,26 and have worse lung function.16 ,34 However, some studies have failed to show a relationship between adherence and clinical outcome score. This is often due to the subjective nature of the adherence measurements used, with inaccurate information skewing the data towards the null hypothesis.16 When objective measurements have been used but no effect on control has been found, it is likely to be because the studies have been underpowered28 ,42 or included very mild asthmatics in whom suboptimal adherence does not have such an effect on clinical outcome.30 One study showed no relationship between preventer adherence and β agonist use,17 possibly reflecting the erratic use of β agonists, and their habitual regular use even in well-controlled asthmatics. Poor disease control is also often associated with poor symptom perception, and β agonist use is often low despite its necessity.

Poor adherence can mislead clinicians into thinking a dose of steroid is ineffective, resulting in unnecessary dose increases and escalation of treatment regimens.

A recent systematic review shows that children with poor adherence have increased healthcare use,43 leading to increased healthcare costs. Poorer adherence rates correlate to increased GP visits,27 ,44 attendances at emergency departments17 ,27 ,45–48 and hospital admissions.27 ,45 ,48 ,49 However, of the nine studies reviewed, only two17 ,27 used electronic monitoring to assess adherence, potentially decreasing the validity of the outcomes.

Why don't children take their inhalers?

There are many reasons why children fail to take their inhalers as prescribed, known as barriers to adherence. These barriers can be divided into intentional and non-intentional. For a child to take their asthma medication, they and their parents have to accept their diagnosis and agree that regular inhalers are necessary to improve the condition. They also have to trust inhaled steroids and have few concerns about their side effects. These intentional factors can be defined as illness perceptions and medication beliefs.50 Studies have shown that adherence is worse in children whose parents have a more negative perception of asthma and doubt the necessity for inhaled steroids, with more concerns about their side effects.32 ,35 ,51 Inhaled steroids have no apparent immediate effect, and it may take up to 6 weeks for a significant clinical improvement to be seen by the patient.38 Therefore, without the appropriate education, patients may decide that inhaled steroids have little or no effect and use them less.

If parents and children have a positive view of asthma, understand the need for regular inhalers and trust inhaled steroids, they are more likely to be motivated to take the medication regularly. In these patients, adherence is usually still suboptimal due to non-intentional factors. When patients and parents are asked about barriers to adherence, commonly reported answers are simply forgetting, being busy and forgetting, trouble incorporating medication into a daily routine, child reaction to medication and being asleep before the evening dose.6 In the paediatric population, responsibility for taking inhalers gradually changes from entirely that of the parents in the younger children, to that of the child in adolescents.52 Younger children have higher rates of adherence21 ,35 suggesting that when the responsibility is that of the parents it is more likely the inhalers will be taken. A study investigating at what age children take responsibility for their daily medications showed that younger children with less responsibility were more adherent to inhaled steroids.51 The complexity of medication regimens also effects adherence, and children with multiple medications at different times are less likely to be able to take them all as prescribed.35 ,51

Non-adherence is likely to be a combination of intentional and non-intentional factors, and in order to improve adherence, both factors need to be addressed. In order for a patient to achieve full adherence, they and their parents need to have a positive view of asthma and inhaled steroids, and have the capability and opportunity to take the medication regularly. Once this has been achieved, the regular behaviour of taking medication can become a habit and ongoing adherence will be more routine.53

What interventions can improve adherence?

Recent meta-analyses have concluded that interventions based on improving education alone have variable effects on adherence rates in children.54 ,55 Some studies have shown that adherence can be improved with educational strategies,11 ,56 ,57 but many have failed to show a significant difference in adherence rates between the two groups.58–60

Behavioural interventions encompass a wide variety of strategies including intensive healthcare support programmes,34 ,61 motivational coaching techniques62 and lifestyle planning.63 The results from trials investigating these interventions are more positive, indicating that adherence and clinical outcome can be improved with behavioural strategies. Recent studies using direct adherence prompts such as alarms64 or text messaging65 have had encouraging results when tested in adult populations.

Unfortunately, the vast majority of studies in the past have used subjective adherence monitoring, and therefore the validity of any results is questionable.11 ,57 ,60–63 ,65 ,66 Objective adherence monitoring with an EMD and structured feedback has been shown to increase adherence rates in children,28 ,42 although these studies used small numbers of patients and were insufficiently powered to detect any difference in clinical outcome. Larger studies in adults have shown increased adherence rates in patients who receive objective adherence feedback.9 By making objective adherence data available to clinicians, the patients who are less adherent can be identified and strategies put in place to increase rates.

Enhancing adherence in the future

Future studies investigating adherence in childhood asthma should strive to record data electronically in order to achieve objective, accurate and reliable results.

Complex interventions involve multiple facets and are successful at improving healthcare measures when carefully devised, and flexible enough to adapt to different population groups.67 Complex interventions are likely to be the most effective way of improving adherence and health outcomes in paediatric asthma as they will be able to address both the intentional and non-intentional barriers to adherence. A successful complex intervention could include an educational or behavioural component to improve illness perceptions and medication beliefs, and thus increase motivation. A second facet to the intervention would involve something practical to improve the capabilities and opportunities for the patient to take their medication.53

The 10-year asthma programme in Finland taught us that effective asthma management involves multiple methods and a multidisciplinary approach with good communication between all healthcare workers.68 Successful management plans involve self-management based on standards agreed by both the patient and the clinician, and an integral part of this should be a review of adherence.

In order for interventions to be successful, as clinicians we must appreciate the complexity of medication taking behaviour and acknowledge our responsibility when prescribing appropriate regimens. Although the term adherence enables an objective measurement of the patient’s medication levels taken, we should aim to achieve concordance, working closely with patients and their families to devise appropriate medication regimens. This process should involve the regular evaluation of both the patients’ adherence and the clinicians’ prescriptions. Changes to one or both of these behaviours can then be implemented in order to achieve successful medication regimens.



  • Contributors RWM, MLE and HEE planned the review. RWM wrote the manuscript, which was reviewed by HEE and MLE, and amendments were made on their recommendation.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.