Accurate reporting of adherence to inhaled therapies in adults with cystic fibrosis: methods to calculate “normative adherence”

Abstract

Background

Preventative inhaled treatments in cystic fibrosis will only be effective in maintaining lung health if used appropriately. An accurate adherence index should therefore reflect treatment effectiveness, but the standard method of reporting adherence, that is, as a percentage of the agreed regimen between clinicians and people with cystic fibrosis, does not account for the appropriateness of the treatment regimen. We describe two different indices of inhaled therapy adherence for adults with cystic fibrosis which take into account effectiveness, that is, “simple” and “sophisticated” normative adherence.

Methods to calculate normative adherence

Denominator adjustment involves fixing a minimum appropriate value based on the recommended therapy given a person’s characteristics. For simple normative adherence, the denominator is determined by the person’s Pseudomonas status. For sophisticated normative adherence, the denominator is determined by the person’s Pseudomonas status and history of pulmonary exacerbations over the previous year. Numerator adjustment involves capping the daily maximum inhaled therapy use at 100% so that medication overuse does not artificially inflate the adherence level.

Three illustrative cases

Case A is an example of inhaled therapy under prescription based on Pseudomonas status resulting in lower simple normative adherence compared to unadjusted adherence. Case B is an example of inhaled therapy under-prescription based on previous exacerbation history resulting in lower sophisticated normative adherence compared to unadjusted adherence and simple normative adherence. Case C is an example of nebulizer overuse exaggerating the magnitude of unadjusted adherence.

Conclusion

Different methods of reporting adherence can result in different magnitudes of adherence. We have proposed two methods of standardizing the calculation of adherence which should better reflect treatment effectiveness. The value of these indices can be tested empirically in clinical trials in which there is careful definition of treatment regimens related to key patient characteristics, alongside accurate measurement of health outcomes.

Keywords:

• cystic fibrosis
• medication adherence
• nebulizers and vaporizers
• epidemiologic methods

Supplementary materials

Further considerations in calculating the “sophisticated” normative adherence index

Extending sophisticated normative adherence: further considerations for the denominator adjustment

The precision of the “sophisticated” adherence index might be increased by also taking into account lung function and the severity of the underlying cystic fibrosis (CF) by considering genotype and pancreatic status. For convenience, we will call this index “extended sophisticated” normative adherence. Whether the additional complexity is of value can potentially be tested empirically and care has been taken to only select adjustment factors that are available in the national CF registries to make such empirical testing feasible in the future.

The additional factors involved in the extended sophisticated normative adherence would allow the index to be sufficiently discriminating to provide guidance as to whether a given adult with CF would require dornase alfa rather than simply assuming that all adults with CF should be prescribed dornase alfa. This has particular relevance because widespread genetic testing is identifying rarer cystic fibrosis transmembrane conductance regulator (CFTR) mutations, leading to CF centers providing care to an increasing population of older and “atypical” cases that would otherwise not be diagnosed as CF.¹ A small group of people with CF have very mild clinical manifestations and near normal lung function even at an sophisticated age.^2–4 In this group of people, there is likely to be less of a consensus about the blanket use of inhaled mucolytic.

Thus, the extended sophisticated normative adherence would potentially identify a group of people who need not necessarily be on any inhaled therapy based on no evidence of Pseudomonas, no history of frequent exacerbations (with the requirement of >14 days of intravenous antibiotics in 1 year as a marker of frequent exacerbation), forced expiratory volume in 1 second (FEV₁) >90%, pancreatic sufficient, and “mild genotype” (at least one class IV–V CFTR mutation⁵).

FEV₁ >90% is accepted by the US CF Foundation as “normal lung function”, whereby long-term dornase alfa is not considered essential.⁶ Pancreatic insufficiency is an independent risk factor for increased FEV₁ decline among people with CF^7–9 and is associated with poorer prognosis.¹ A potential disadvantage of relying solely on the pancreatic status to identify milder phenotype is that people with mild phenotype who were initially pancreatic sufficient may eventually become pancreatic insufficient after a series of episodes of pancreatitis.¹⁰ Although there is significant phenotypic variability for each class of CFTR mutation, the relationship between pancreatic status and genotype is more robust and the group with at least one class IV–V CFTR mutation does tend to have milder lung disease.^2,5,11,12 Therefore, the genotype is useful in supplementing the information provided by pancreatic status in terms of confirming that an individual has a milder phenotype.

Figure S1 summarizes the rubric for combining the different prognostic factors used to determine the required maintenance inhaled therapy for this form of sophisticated normative adherence index.

Taking into account incomplete doses: numerator adjustment for adherence levels calculated from I-neb^® data

The I-neb^® records four different readings for each nebulizer dose depending on treatment completeness: “full” = full nebulizer dose delivered; “12.5%–100%” = treatment taken but incomplete dose; “<12.5%” = treatment attempted but unlikely to receive any; and “none” = I-neb^® switched on but no treatment taken.¹³ To ensure that the number of treatments is correctly calculated, a ”full” dose is counted as “1 dose”, “12.5%–100%” is counted as “½ dose” while “<12.5%” and “none” doses are counted as 0.

Taking into account doses taken after midnight: numerator adjustment for irregular lifestyles

When a limit of 100% adherence per day is being used as part of numerator adjustment, it can be informative to recognize that many young people will go to bed after midnight. It is not uncommon for these “night owls” to use their inhaled therapy after midnight (eg, after returning from a night out). For example, a person may use his inhaled antibiotic at 10 am and take the final dose of the day just before bed, which may on occasions be 1 or 2 am. Let us say the second lot of inhaled therapy (this time an inhaled antibiotic and inhaled dornase alfa) was used around 1 am the next morning. He/she woke up around 10 am to start his/her new day and used his/her morning inhaled antibiotic. Finally, he/she used his/her second lot of inhaled therapy (inhaled antibiotic and inhaled dornase alfa) at around 11 pm. The unadjusted adherence over the 2 days would be 100% (six nebulizers used out of six prescribed). Capping the daily maximum at 100% using a rigid midnight-to-midnight day would result in one nebulizer counted for the first day and three nebulizers counted for the second day (ie, adherence over the 2 days would only be 67%). Counting a day as starting at 5 am and ending at 4.59 am is a pragmatic solution to this problem, since it is likely that on most occasions a person would go to bed by 5 am and would wake up for the day by 4 pm.¹⁴

Therefore, to prevent intermittent late doses pushing some days over 100% and leaving other days under 100%, which would lead to a lower overall adherence level due to capping the maximum daily inhaled therapy use at 100%, “a day” should be considered to start at 5 am.

Taking into account dose spacing: numerator adjustment for inhaled antibiotics

Another factor that could be considered in ensuring that the adherence index most accurately reflects medication effectiveness is to consider dose spacing.

The common inhaled antibiotic therapies (colistimethate sodium and tobramycin) in CF should be used twice daily, that is, every 12 hours, while inhaled aztreonam lysine dosing is thrice daily. Inhaled antibiotic doses used too close together may not be as beneficial as doses used at the recommended intervals. The UK CF Trust recommends a minimum interval of at least 6 hours for inhaled colistimethate sodium and tobramycin.¹⁵ Therefore, the numerator adjustment could exclude the inhaled antibiotic doses that were used <6 hours after an initial dose.

Taking into account device dose delivery characteristics: numerator adjustment for nebulized tobramycin via I-neb^®

Nebulized tobramycin via the I-neb requires two separate nebulizations to complete a single dose due to the size of the chamber. Thus, the number of nebulizations per day will differ depending on whether the patient is on a mucolytic and tobramycin or mucolytic and colistimethate sodium. A person would take one nebulization for the dornase alfa and one nebulization for the colistimethate sodium in the morning and just one nebulization for the colistimethate sodium in the evening with a target of three nebulizations per day. If the patient was taking dornase alfa and tobramycin via the I-neb^®, the morning nebulization target would be one nebulization for dornase alfa and the patient would need to use the nebulizer twice to deliver the full dose of tobramycin. Hence, the patient would have a target of five nebulizations per day.

A relatively common scenario is for people to use both an inhaled mucolytic and inhaled antibiotic within a “treatment session”, but miss their other inhaled antibiotic for the day. If the inhaled antibiotic is colistimethate sodium, the adherence would be 67% (two out of three nebulizers used). However, if the inhaled antibiotic is tobramycin solution, which requires two separate nebulizations for a complete dose via the I-neb, the adherence would only be 60% (three out of five nebulizers used). This discrepancy does not arise with other types of nebulizers, for example, the eFlow Rapid^®, which does not require the tobramycin solution to be nebulized twice for a complete dose. Missing a dose of colistimethate sodium via the I-neb^® should carry the same weight as missing a dose of tobramycin. Therefore, the numerator adjustment counts each nebulization of tobramycin solution via the I-neb^® as “½ dose”, so that the complete dose (two nebulizations) would count as “1”. This allows the daily denominator to stay at “3” for those on both inhaled antibiotic and inhaled dornase alfa, thus avoiding the discrepancy between missing a dose of colistimethate sodium versus missing a dose of tobramycin.

Figure S1 The required maintenance inhaled therapy based on a range of prognostic factors used to decide the minimum denominator for the “extended sophisticated” normative adherence.

Notes: ^ψPseudomonas status as defined by the Leeds definition.¹⁶ People with intermittent Pseudomonas should be on inhaled antibiotics in addition to inhaled mucolytic for 1 month or 3 months depending on the antibiotic regime when Pseudomonas is newly detected. ^ΩGenotype status as defined by international consensus.⁵ “Mild genotype” is defined by the presence of at least one class IV–V CFTR mutation.

Abbreviations: CFTR, cystic fibrosis transmembrane conductance regulator; FEV₁, forced expiratory volume in 1 second; IV, intravenous.

References

• HooZHWildmanMJTeareMDExploration of the impact of ‘mild phenotypes’ on median age at death in the U.K. CF registryRespir Med2014108571672124675238
   PubMed Web of Science ®Google Scholar
• McKoneEFEmersonSSEdwardsKLAitkenMLEffect of genotype on phenotype and mortality in cystic fibrosis: a retrospective cohort studyLancet200336193701671167612767731
   PubMed Web of Science ®Google Scholar
• RodmanDMPolisJMHeltsheSLLate diagnosis defines a unique population of long-term survivors of cystic fibrosisAm J Respir Crit Care Med2005171662162615591474
   PubMed Web of Science ®Google Scholar
• SimmondsNJCullinanPHodsonMEGrowing old with cystic fibrosis – the characteristics of long-term survivors of cystic fibrosisRespir Med2009103462963519022643
   PubMed Web of Science ®Google Scholar
• CastellaniCCuppensHMacekMJrConsensus on the use and interpretation of cystic fibrosis mutation analysis in clinical practiceJ Cyst Fibros20087317919618456578
   PubMed Web of Science ®Google Scholar
• MogayzelPJJrNaureckasETRobinsonKACystic fibrosis pulmonary guidelines. Chronic medications for maintenance of lung healthAm J Respir Crit Care Med2013187768068923540878
   PubMed Web of Science ®Google Scholar
• KonstanMWMorganWJButlerSMRisk factors for rate of decline in forced expiratory volume in one second in children and adolescents with cystic fibrosisJ Pediatr2007151213413917643762
   PubMed Web of Science ®Google Scholar
• KeremEVivianiLZolinAFactors associated with FEV1 decline in cystic fibrosis: analysis of the ECFS patient registryEur Respir J201443112513323598952
   PubMed Web of Science ®Google Scholar
• KonstanMWWagenerJSVanDevanterDRRisk factors for rate of decline in FEV1 in adults with cystic fibrosisJ Cyst Fibros201211540541122561369
   PubMed Web of Science ®Google Scholar
• AugartenABen TovAMadgarIThe changing face of the exocrine pancreas in cystic fibrosis: the correlation between pancreatic status, pancreatitis and cystic fibrosis genotypeEur J Gastroenterol Hepatol200820316416818301294
   PubMed Web of Science ®Google Scholar
• No authors listedCorrelation between genotype and phenotype in patients with cystic fibrosis. The Cystic Fibrosis Genotype-Phenotype ConsortiumN Engl J Med199332918130813138166795
   PubMed Web of Science ®Google Scholar
• McKoneEFGossCHAitkenMLCFTR genotype as a predictor of prognosis in cystic fibrosisChest200613051441144717099022
   PubMed Web of Science ®Google Scholar
• NikanderKArhedenLDenyerJCobosNParents’ adherence with nebulizer treatment of their children when using an adaptive aerosol delivery (AAD) systemJ Aerosol Med200316327328114572325
   PubMedGoogle Scholar
• HooZHGardnerBCurleyRWildmanMJPart III: constructing a process behaviour (XmR) chart with I-neb adherence data for people with cystic fibrosisJ Im Sci201312112
   Google Scholar
• UK Cystic Fibrosis Trust Antibiotic Working GroupAntibiotic treatment for cystic fibrosis52009 [cited January 31, 2016]. Available from: https://www.cysticfibrosis.org.uk/~/media/documents/the-work-we-do/care/consensus-documents/antibiotic-treatment-cystic-fibrosis-may-09.ashx?la=enAccessed January 31, 2016
   Google Scholar
• LeeTWBrownleeKGConwaySPDentonMLittlewoodJMEvaluation of a new definition for chronic Pseudomonas aeruginosa infection in cystic fibrosis patientsJ Cyst Fibros200321293415463843
   PubMedGoogle Scholar

Acknowledgments

This report presents independent research funded by the NIHR under its Grants for Applied Research Programme (Grant Reference Number RP-PG-1212-20015) and a Doctoral Research Fellowship (Zhe Hui Hoo, Award Identifier DRF-2014-07-092). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, Medical Research Council (MRC), Central Commissioning Facility (CCF), NIHR Evaluation, Trials and Studies Coordinating Centre (NETSCC), the Programme Grants for Applied Research Programme, or the Department of Health. Rachael Curley received support from Zambon and Philips Respironics for a parallel research study monitoring inhaled adherence. Martin Wildman received funding from Zambon and support from Philips Respironics for the same study. This has not had any direct influence on this submitted paper. In addition, Martin Wildman has worked with Pari to carry out studies using the chipped E-flow (e-track).

Disclosure

The authors report no conflicts of interest in this work.