Key learning points

  • Low adherence with medication may result in people with cystic fibrosis presenting with declining lung function
  • Self-reported adherence tends to overestimate adherence levels, although some people do underestimate their adherence
  • Low adherence in cystic fibrosis is often multifactorial and is not just about treatment burden
  • Habit appears to have an important role in sustaining adherence among adults with cystic fibrosis so ways to encourage habit formation may be used to support adherence


Cystic fibrosis (CF) is a long-term condition predominantly affecting the lungs and digestive system in which the daily use of nebulised medications is crucial to maintain health. There is often a sense of pessimism regarding the potential to improve low adherence with medications because previously there are no adherence interventions with proven efficacy for people with CF.1 Previously, the largest adherence intervention trial among adolescents with CF was carried out in the US and published in 2019 – the trial found no difference between problem solving and education intervention versus usual care in improving adherence rates.2

However, it is important to note that regardless of the potential to alter adherence, understanding adherence levels among people with CF is still critical to ensure efficient clinical management. A common scenario during clinic is for a person with CF to present with declining lung function. This typically prompts various investigations to look for complications such as allergic bronchopulmonary aspergillosis or a new diagnosis of CF-related diabetes, yet with median adherence rates of only 35–50% among adults with CF,3,4 the cause for worsening lung health is often low medication adherence. Another common response to declining lung function is the escalation of twice-daily nebulised tobramycin to thrice-daily nebulised aztreonam, but treatment escalation would not address the problem if lung function decline is a result of tobramycin not being used in the first place. Conversely, it is not uncommon for declining lung function to be attributed to low adherence when actually adherence is already optimal,3 and incorrect assumptions can result in clinicians missing opportunities to escalate treatment or investigate for complications. Understanding and investigating problems with non-adherence will help to avoid lamp post syndrome medicine (a type of observational bias that occurs when people only search for something where it is easiest to look, so-called because of the story of a man who drops his keys on a dark street but walks two blocks away to look for the keys underneath a lamp post because that is where the light is),5 and so improve the health outcomes of people with CF.

Ways to measure adherence

To understand adherence levels among people with CF, it is important to use an appropriate method of capturing adherence data. The two most common methods in the literature are self-report (asking respondents to characterise their medication-taking behaviour) and pharmacy refill (measuring how much of the prescribed medications are actually collected) but both methods have their limitations. Self-report tends to overestimate adherence,3,6 but both under- and over-estimation of self-reported adherence levels occur without any clear predictors for the direction of errors.3 A substantial minority of people underestimate their adherence,3 which could potentially result in poor prescribing practices if clinicians were to act on those self-reported measures.

Pharmacy refill may be somewhat more reliable than self-report, in that it represents the maximum amount of medication a person can take.7 However, not all collected medications will be used correctly and some medications may not even be used at all. One study evaluating adherence with ivacaftor among people with CF found higher adherence levels as measured by pharmacy refill (median 84%) compared with electronic data capture (EDC, median 61%).8 Another limitation is the inability of pharmacy refill to capture important granularity of behaviours involved in using preventive inhaled therapies.

EDC is generally regarded as the ‘gold standard’ method to capture adherence data because of its accuracy.9 In CF, data-logging nebulisers such as I-neb® and eTrack® provide date- and time-stamped data for every dose of nebulised medication used. Although they are more costly than devices without data-logging capability, EDC devices could potentially bring about cost savings if they are used effectively to provide feedback and improve adherence.10 Another advantage of EDC is that it provides a continuous stream of rich quantitative data that allow adherence patterns to be studied in greater detail. For example, a study using I-neb® found that adolescents with CF are most adherent during weekdays of school term-time.11 This is when adolescents with CF are likely to be busiest and therefore might be suspected to be most susceptible to lapses in adherence due to the time-consuming and burdensome nature of treatment. This finding challenges the presumption that treatment burden is the main driver of low adherence in CF and demonstrates the potential insight that can be derived by studying adherence using EDC. (Other data also indicate that treatment burden is not the only cause of low adherence in CF. One study actually found increasing levels of adherence with increasing numbers of prescribed pulmonary medications.4 Older adults tend to have more complex treatment regimens than younger adults, but older adults may have higher adherence levels.4 Ivacaftor only involves swallowing tablets twice daily and is far less burdensome than inhaled therapies but EDC has shown the adherence level is still only about 60%.8 In fact, low levels of adherence are pervasive across all types of long-term conditions, regardless of treatment modalities or burden.12,13)

Factors influencing adherence

Most of the studies investigating adherence determinants in CF have eschewed the use of objective adherence data and explicit psychological theory; and have instead relied on subjective recall and focusing on the practical barriers/facilitators to adherence. These types of studies are prone to both recall and social desirability bias. Studies using data-prompted interview guided by a comprehensive psychological model such as COM-B – which posits that adequate capability (C), opportunity (O) and motivation (M) are required for a behaviour (B) to occur – have found that a wide range of factors potentially influence adherence.14 Low adherence in CF is often multifactorial and is not just about treatment burden.

Research into medication adherence in other long-term conditions has highlighted the potential importance of automatic processes such as ‘habit’, that is, automatically experiencing an urge to take a treatment in certain settings. Automatic control of habitual behaviour reduces the dependence on conscious attention, so that habitual behaviours should persist even if conscious motivation wanes.15 Habitual behaviours are also cognitively efficient and can free up mental resources for other tasks,15 for example, someone using a nebuliser habitually could have more cognitive resources to pay attention to the treatment of CF-related diabetes. Habit has been found to better predict adherence than conscious motivational factors in long-term conditions,16–19 and a meta-analysis of 771 adherence trials found interventions that included habit formation were more effective than those that did not.20 There is emerging evidence that habit has an important role in sustaining adherence among adults with CF;21,22 and investigation of habit formation as a potential intervention to support adherence in CF should be seen as a priority.

It should be noted that although adherence data captured by EDC is highly accurate, the effective adherence level can still be overestimated if efficacious treatments are inadequately prescribed. (Effective adherence is important to consider because the fundamental purpose for supporting nebuliser adherence among adults with CF is to improve health outcomes. The ABC taxonomy for medication adherence recommends that adherence definitions should account for deviation that adversely influences the effectiveness of a treatment regimen.23 The Respiratory Effectiveness Group also emphasises that adherence levels should reflect treatment effectiveness.24) Adherence is typically reported as a percentage. When percentage adherence increases, improvement in effectiveness, that is, an increase in the numerator (number of doses taken) is often assumed. Yet percentage adherence also increases when the denominator (target number of doses to be taken) decreases, which is likely to represent a decrease in effectiveness. In general, inadequate prescription of efficacious treatment (sometimes termed clinical inertia) is the second biggest cause of ineffective treatment after low adherence.25 In CF, the potential for clinical inertia to result in exaggerated effective adherence levels is particularly relevant; because treatment burden is widely presumed to be the main barrier to adherence and treatment rationalisation is often employed as a strategy to support adherence, however, only around two-thirds of people with CF are prescribed the recommended inhaled therapies.26

Another reason for overestimating effective adherence levels is a sampling frame that ignores the group of people who are unwilling to share their adherence data (and who are likely to have lower than average adherence) or refuse to use any treatment. Since the CF literature tends to only report agreed adherence from convenience samples without accounting for the appropriateness of treatment prescription or people with missing adherence data, the actual effective adherence among adults with CF is likely to be even lower than published. This perspective is important in highlighting the scale of the problem that is facing people with CF.

In response, the CFHealthHub group, which is a UK collaboration, is working towards developing, evaluating and implementing effective adherence interventions in CF. First, the CFHH RCT (ISRCTN5504164), a randomised controlled trial, was carried out among 608 people with CF in 19 centres to develop and evaluate a complex adherence intervention combining a display of real-time adherence data with a comprehensive suite of behaviour change techniques. The trial has recently been completed and has demonstrated significantly increased adherence sustained to 80 weeks which was associated with reduced effort with treatment-taking and reduced subjective burden alongside a significant increase in habit (Eur Respir J, draft in review).

Second, the CFHH improvement collaborative (ISRCTN14464661) is set up as a learning health system, in which technical, behaviour change and implementation science approaches are explicitly used to learn and improve. Three centres are involved in the improvement collaborative at present, but 10 more adult CF centres across the UK will join following the conclusion of the CFHH RCT. The improvement collaborative aims to embed the use of objective adherence data and implement effective behaviour change interventions in routine clinical practice. It is hoped that these projects will extend our understanding of adherence and pave the way for sustained adherence among adults with CF.

Dr Martin Wildman, Consultant Respiratory Medicine & Adult CF, Sheffield Adult CF Centre and Dr Hui Hoo, Consultant Respiratory Medicine & Adult CF, Sheffield Adult CF Centre

This project was initiated and funded by Teva Respiratory. Teva have had no influence over content. Topics and content have been selected and written by independent experts.