Compliance with an oral asthma medication: a ... - Respiratory Medicine

RESPIRATORY MEDICINE (2000) 94, 852±858 doi:10.1053/rmed.2000.0813, available online at http://www.idealibrary.com on

Compliance with an oral asthma medication: a pilot study using an electronic monitoring device K. F. CHUNG*

AND

I. NAYA{

{

National Heart and Lung Institute, Imperial College School of Medicine & Royal Brompton & Harefield Hospital, London and{ AstraZeneca, Macclesfield, U.K.

Compliance with prescribed asthma medication is commonly estimated from tablet counts for oral medications and canister weights for inhaled medications. Recently, electronic medication monitoring devices, developed to evaluate numerical compliance as well as drug use patterns, were used to assess compliance with inhaled steroids and b2agonists. This was the ®rst study to electronically assess compliance with an oral asthma medication. Fifty-seven asthmatic patients, stable on inhaled b2-agonists only with a mean FEV1 of 77% predicted ( 13%, SD) began 12 weeks of treatment with za®rlukast 20 mg twice daily. The monitoring device, an electronic TrackCapTM, recorded the date and time on each occasion that patients removed and replaced their medication bottle caps. Patients were told that compliance would be assessed as part of the study, but patients were not told about the speci®cs of the TrackCapTM. Compliance was de®ned: 1. as the number of TrackCapTM events per number of prescribed tablets; and 2. as the dierence between number of tablets dispensed and number returned per number prescribed. Adherence was de®ned as the number of days with two TrackCapTM events at least 8 h apart per the total number of days' dosing. Forty-seven patients completed the study with a median compliance of 89% (mean, 80%) and a median adherence of 71% (mean, 64%) as measured by TrackCapTM events. Compliance as estimated from return-tablet count was slightly higher (median, 92%). High rates of compliance were maintained throughout the trial. These results show that compliance with and adherence to a treatment of an oral, twice-daily, maintenance asthma medication, such as za®rlukast, is high. Key words: compliance; asthma; electronic TrackCapTM; za®rlukast; leukotriene-receptor antagonist. RESPIR. MED. (2000) 94, 852±858

Introduction Patient compliance with prescribed treatments for chronic diseases, such as asthma, is an essential aspect of disease management. Yet studies have shown that only one third of all patients and 0 to less than 50% of asthma patients take their medications as prescribed (1±7). In patients with asthma, treatment non-compliance is thought to play a major role in the unacceptably high morbidity and mortality associated with the disease (8±11). To what extent non-compliance is associated with a lack of disease comprehension or with diculty in using medication dispensing devices is unclear; however, both appear to aect a patient's inclination to either adhere to or disregard Received 17 January 2000 and accepted in revised from 25 February 2000 Financial support: This trial was supported by a research grant from AstraZeneca, Maccles®eld, U.K. Correspondence should be addressed to: Kian Fan Chung MD, FRCP Address: Imperial College School of Medicine at The National Heart and Lung Institute Dovehouse Street London SW3 6LY, U.K. Fax: 44 171 351 8126.

0954-6111/00/090852+07 $35?00/0

# 2000 HARCOURT PUBLISHERS LTD

treatment advice. In fact, compliance with asthma medications is related to several factors including: patient understanding of treatment rationale; simplicity of the treatment regimen; medication eectiveness and ease of use; social support; and a good physician±patient relationship (1,3, 12). In adolescents, who are particularly non-compliant, compliance is more likely to be enhanced when treatment plans additionally address steroid phobias, peer pressure, smoking, and other lifestyle issues (13). However de®ned, compliance is dicult to measure. Doing so, however, is a critical step in evaluating whether or not strategies aimed at improving asthma medication compliance are succeeding. Currently, both direct and indirect methods are used (14). Direct measures include assay of blood, urine, or saliva to con®rm drug use and direct observation to con®rm that dispensing devices or metered-dose inhalers (MDIs) are properly used. The value of these objective assessments is oset by several disadvantages (14). The assay approach is limited by the small number of assays available for asthma medications and by the invasiveness of tissue sampling. Direct observations generally occur only at clinic visits and, although patients may demonstrate adequate skills with MDIs while under # 2000 HARCOURT PUBLISHERS LTD

COMPLIANCE WITH AN ORAL ASTHMA MEDICATION 853 observation, they may not consistently apply such skills in all environments. Neither approach provides information regarding the pattern of drug use (14). Indirect measures include clinical judgement, self-reported asthma diaries and medication measurements (tablet counts, canister weights, etc). These methods tend to overestimate compliance and only infer drug use with varying degrees of reliability depending on the individuals involved. The introduction of electronic monitors that objectively record dates and times of bottle openings or MDI actuationsÐknown as medication-use eventsÐprovides an advantage over other indirect methods because researchers now can study patterns of drug use (14). Noteworthy among the compliance literature is the ®nding that patients with asthma tend to be more compliant with their oral asthma medication (e.g., theophylline) than with their inhaled asthma medication (e.g., corticosteroids and cromolyn sodium) (12). Oral leukotriene-receptor antagonists, such as za®rlukast, may provide a signi®cant alternative for dealing with poor compliance and may ultimately improve asthma outcomes (13,15). Orally administered za®rlukast has been shown to reduce asthma symptoms and rescue use of bronchodilators, improve lung function, improve quality of life measurements, and decrease exacerbations when administered twice a day (16±18). Furthermore, early evaluation of patient compliance with za®rlukast therapy is promising. In a study evaluating the long-term ecacy and safety of za®rlukast, the compliance rate based on tablet counts was 94?5% at 52 weeks of treatment (19). Because tablet count methods tend to overestimate compliance and do not provide insight into drug use patterns, a pilot compliance study, as reported here, was conducted with za®rlukast to measure long-term compliance electronically and to compare results with compliance calculated from tablet counts for individual patients.

Methods and materials PATIENTS Male and female patients, 18±55 years of age, with a history of asthma were sought for enrollment. Asthma history was de®ned as previous patient response to standard asthma treatment, episodic wheezing, or changes in lung function over short periods of time. Patients who were using any asthma medication, other than as-needed, short-acting b2agonists, were excluded from screening. Patients eligible to receive trial treatment were those who, during the screening period, had a 1-sec forced expiratory volume (FEV1) of 60% predicted; a15% improvement in peak expiratory ¯ow (PEF) or FEV1 after a 400-mg inhaled dose of salbutamol; and symptomatic reversible airways obstruction that required additional controller asthma therapy, as determined from screening diary cards. Major reasons for exclusion included other signi®cant respiratory disease; recent hospitalization for asthma; recent history of respiratory tract infection; seasonal asthma; use of corticosteroids, cromones, or long-acting

2 -agonists in the weeks that preceded the trial; recent participation in another clinical trial; and any clinically signi®cant medical condition that would interfere with trial assessments or increase the patient's risk. The study was conducted in accordance with the Declaration of Helsinki and was approved by the designated ethics committee at each centre. Written informed consent was obtained after patients received a description of the trial's purpose and procedures. The informed consent form mentioned that compliance, pulmonary function and safety would be measured. Patients, however, were not informed that compliance was under electronic surveillance so as not to introduce an unacceptable bias in favour of compliance (5). Standards of ethics were considered maintained because patients were not subjected to additional risk, and patients' rights and welfare were neither violated nor aected.

STUDY DESIGN This study was a 15-week, open-label, non-comparative trial and the ®rst to de®ne compliance for za®rlukast according to a dosing schedule. An initial screening period of 2±3 weeks was followed by a 12-week treatment period during which patients took za®rlukast 20 mg twice a day. The purpose of the screening period was to ®nd out if patients had reversible airway obstruction that made them suitable candidates for a stepwise increase in asthma therapy. At the start of the treatment period, each patient received 56 tablets of za®rlukast, enough for 3 weeks, with one week's supply to spare. Tablets were dispensed in screw-top bottles ®tted with a TrackCapTM medication event monitoring system (MEMS) device (APREX Corporation, Fremont, California, U.S.A.), that recorded the date and time on each occasion when the cap was removed and replaced. Patients were scheduled to return to the clinic every 3 weeks for four more visits. At three of those visits, patients exchanged bottles with unused tablets for newly stocked bottles of 56 tablets; at the fourth visit, no new supplies were issued. The following dosing instructions were on the bottle and outer cartons: `Take one tablet in the morning and one tablet in the evening approximately 12 h apart. Do not take the tablets at mealtimes.' Patients were also instructed to remove one tablet at a time and immediately replace the cap because the tablets were sensitive to moisture. A desiccant capsule, found in each bottle, visually reinforced the need to replace the cap as directed. Patients were not required to keep asthma diaries during the treatment period, and tablet counts were neither assessed nor discussed in the patient's presence. Each removal of the TrackCapTM was presumed to indicate a single medication-use event. The device, however, was programmed to recognize, but not accumulate, multiple openings that occurred within 1 min of each other. If the cap was left o for 15 min or more, the device recorded one additional event. To prevent extraneous results, study personnel were asked not to open the bottles before, during, or after the study period.

854

K. F. CHUNG AND I. NAYA

Patients were assessed at each clinic visit for lung function [peak expiratory ¯ow rate (PEF) and forced expiratory volume in 1 sec (FEV1)] and adverse events. Patients who withdrew underwent a complete physical examination and a standard battery of laboratory and pulmonary function tests at the time of withdrawal.

ANALYSIS

32 men and 25 women, had a mean age of 29 years, and they had a mixed socioeconomic background. At entry, mean % predicted FEV1 was 77%, and duration of asthma ranged from 10 months to 42 years. Additional patient characteristics are summarized in Table 1. Of the 57 patients who received za®rlukast, 47 completed the prescribed 12 weeks of treatment. Among patients withdrawn were three who did not comply with the

Percent TrackCapTM compliance was de®ned as follows: % Compliance 100

Number of TrackCapTM events Number of prescribed tablets

Compliance also was estimated from returned tablet counts, with percent compliance de®ned as: % Compliance 100

Number of dispensed tablets minus number of returned tablets Number of prescribed tablets

TrackCapTM adherence was the degree to which patients followed dosing instructions precisely on a daily basis and was de®ned as follows: % Adherence 100

Number of days with 2 TrackCapTM events at least 8 h apart Total number of days0 dosing

In addition, days were classi®ed as days with two events 58 h apart (insucient interval adherence), days with only one event (undercompliance), days with no event (no compliance), and days with more than two events (overcompliance). A day was de®ned as the 24-h period from 03?00 h on one day to 02?59 h on the next day. For each category, the total number of days that a patient was classi®ed as such was calculated and expressed as a percentage of the total number of days the patient was on trial treatment. Summaries of compliance and adherence were based on data from patients who completed the entire 12 weeks of treatment. In the evaluation of TrackCapTM data, only events that were recorded during the treatment period were considered, i.e., events that occurred before tablets were dispensed or after containers were returned were discounted. Initial plans to calculate 95% con®dence intervals for mean compliance and adherence were not applicable after a test for normality (Shapiro±Wilk test) showed that compliance and adherence data were not normally distributed. Therefore, both means and medians are reported. To evaluate consistency of compliance over time, as well as pattern of drug use relative to clinic visits, mean compliance was calculated weekly for all patients for whom data were available during that week. Patients who withdrew were included in this analysis because they were more likely to represent a non-compliant population.

Results DEMOGRAPHY A total of 75 patients from ®ve sites underwent screening for possible study entry; of those, 57 met entry criteria and began treatment with za®rlukast. The study population,

protocol, two who did not return for follow-up visits, two who withdrew consent, one who had an adverse event (mild headache), and one who was erroneously enrolled.

TRACKCAPTM ADHERENCE AND COMPLIANCE As calculated from TrackCapTM events, median adherence was 71% (mean, 64%) of the days of dosing (lower quartile, 51%; upper quartile, 81%), and median compliance was 89% (mean, 80%) (Tables 2 and 3). The distribution of percent adherence and compliance values are shown in Figs 1 and 2. Sixty-six percent of the patients had good compliance according to TrackCapTM measurements (took 80% of the prescribed doses). On days when patients took exactly two tablets, the mean time between doses was 12 h 34 min (range, 17 min to 21 h 1 min), with tablets TABLE 1. Demographic and clinical characteristics at entry Characteristic Men/Women, n (%) Age (years) Mean (SD) Range Race, n Caucasian/Asian Mean % predicted FEV1 (SD) Mean % FEV1 reversibility (SD) Mean % PEF reversibility (SD) Mean years with asthma (SD) SD:

(n=57) 32 (56)/25 (44) 29 (9) 18 to 55 56/1 77% (13) 26% (10) 20% (7) 14 (10)

standard deviation; FEV1: forced expiratory volume in 1 sec; PEF: peak expiratory ¯ow.

COMPLIANCE WITH AN ORAL ASTHMA MEDICATION 855 TABLE 2. Summary data for compliance, n=47

TrackCapTM compliance* Tablet count compliance{

Mean (SD) %

Lower quartile %

Median %

Upper quartile %

80 (24) 89 (11)

74 84

89 92

96 96

*Number of events divided by number of prescribed tablets6100.{ Number of dispensed tablets minus returned tablets divided by the number of prescribed tablets6100. TABLE 3. Summary data for adherence, n=47 Mean (SD) % of days Full adherence (2 events 8 h apart) Insucient interval adherence (2 events 58 h apart) Undercompliance (1 event/day) No compliance (0 events/day) Overcompliance (>2 events/day)

FIG. 1. Distribution of percentages of days with full adherence to za®rlukast therapy as determined by TrackCapTM. taken 8±16 h apart on 86% of treatment days (total treatment days, 2624). Patients had one TrackCapTM event on a median of 19% of the days, and had insucient interval adherence on a median of 2% days (Table 3). The median percent of days when patients took no tablets was 1%. Patients were overcompliant (>two tablets day71) on a median of 2% of days.

COMPLIANCE ESTIMATED FROM TABLET COUNT On the basis of tablet counts, median compliance was 92% (mean, 89%), which was slightly higher than compliance calculated from TrackCapTM events (Table 2). For 39 of 47 patients (83%), the dierences between values derived from

64 3 20 10 2

(26) (4) (14) (21) (2)

Lower quartile % of days

Median % of days

Upper quartile % of days

51 1 10 0 0

71 2 19 1 2

81 4 25 8 4

FIG. 2. Distribution of percentages of prescribed doses of za®rlukast taken (compliance) as determined by TrackCapTM. the two methods were marginal (within +15%), suggesting that patients removed multiple tablets at a time infrequently (Fig. 3). For eight patients, however, compliance based on tablet counts was greater by 20% or more, indicating that these patients did not remove trial medication as instructed. The pattern of TrackCapTM events suggested that one patient removed up to a week's supply of tablets at each opening. Two patients (4%), recorded multiple events (up to 37 bottle openings within a 2-min period) before clinic visits, which suggested that these patients suspected compliance monitoring and attempted to conceal poor compliance.

WEEKLY COMPLIANCE Three patients did not return the TrackCapTM device, but all other patients provided at least a week of TrackCapTM event

856


FIG. 3. TrackCapTM observed compliance. vs. tablet count compliance with za®rlukast therapy.

data. The weekly mean data are summarized in Fig. 4. The mean level of compliance over time remained consistently high from week 1 through week 12 of treatment. Clinic visits did not appear to have a great eect on compliance.

OTHER ASSESSMENTS Za®rlukast was well tolerated throughout the trial. The most frequently reported adverse events were non-treatment-related infections (primarily respiratory infections) in 11 patients (19%) and headache in 10 patients (18%). Mean FEV1 remained unchanged but mean PEF increased 8% from the end of screening.

Discussion This was the ®rst study to evaluate compliance with an oral asthma medication using a medication event monitoring system. As assessed by the electronic TrackCapTM device, compliance with twice-daily oral za®rlukast was high (median, 89%). These results con®rm the high levels of compliance reported for za®rlukast in clinical trials, as estimated from returned-tablet counts (19). Also, 66% of patients achieved a good level of compliance. Importantly, through 12 weeks of treatment, levels of compliance remained consistently high. Although median adherence in this study was not as high as median compliance, it was also high at 71%. The dierence between these assessments was primarily associated with days when one tablet was taken. Days of no compliance accounted for only a median of 1% of the days. Overall, then, patients took za®rlukast according to dosing instructions for most of the 12-week treatment period; on most other days, patients took at least half the daily dose. Adherence with za®rlukast was more than ®ve times greater than adherence measured electronically (14%) for a fourtimes-daily regimen of inhaled nedocromil (4).

FIG. 4. Weekly median TrackCapTM compliance with za®rlukast therapy and weekly percentages of patients with TrackCapTM compliance 80% (numbers of patients in parentheses). Not surprising was the fact that mean compliance calculated from the returned-tablet counts (89%) was greater than that calculated from TrackCapTM events (80%). This phenomenon of overestimation of compliance by tablet count may be attributed to accidental loss of medication, removal of multiple tablets at bottle openings, or deliberate disposal of tablets to disguise underusage. A similar corollary is documented for inhaled asthma medication in that residual MDI canister weights and daily use records also overestimate compliance (2,4,20,21). In this study, the dierence in compliance as calculated by the TrackCapTM and tablet-count methods was minor and mostly could be attributed to eight patients who clearly did not follow speci®c dosing instructions. However, just as it was possible that some patients may have removed tablets and not taken them, it was possible that some patients, despite their failure to comply with instructions regarding tablet removal, still took the medication according to the prescribed regimen. The overall conclusions that can be drawn from this study are limited in that, without a comparator, we could not assess how well these patients might have complied with other asthma treatments. We cannot exclude the possibility that these patients were simply a more compliant group. However, mean compliance according to tablet counts was 94?5% in a long-term za®rlukast trial in North America (19), which suggests that the patients studied in this trial did not represent a highly compliant group. The patients included in our study were not previously following a regular regimen of asthma therapy. It is unclear if patients who were previously on a regular regimen might have had a higher compliance. A compliance rate of 79%, which was similar to that measured in our study, was reported for twice-daily theophylline when medical records from an health maintenance organization population were retrospectively compared with prescription re®ll records (12). The theophylline study had the disadvantage of not being

COMPLIANCE WITH AN ORAL ASTHMA MEDICATION 857 able to assess daily compliance but had the advantage of not introducing the potential for bias related to patient selection and study participation. In our study, compliance rates may have been in¯uenced because patients knew they were in a trial and had obligations to return at set intervals. In several compliance studies that required follow-up visits, patients demonstrated a variety of behaviours related to drug administration, e.g., increasing medication use around clinic visits, taking only half the recommended dose for the trial duration, or taking medication as directed for selfdetermined blocks of time (5,22,23). None of these behaviours occurred in this trial, which lends credibility to the rate of compliance and adherence achieved. Patterns of multiple-grouped TrackCapTM events did suggest that some patients realized the purpose of the device. Among the various direct and indirect methods currently used to assess compliance, none does so perfectly (14). Direct methods are ultimately inconvenient, and indirect methods, with inference of drug use, are problematic in determining the extent of full compliance and accounting for patient behaviours. The precedence for use of electronic methods was previously established in studies that evaluated compliance in patients with hypertension (24) and epilepsy (25). In addition to achieving adherence rates similar to the rate achieved in this study, those studies also revealed that adherence (full compliance) was inversely proportional to frequency of dosing. Eorts to prevent a bias in favour of compliance in this studyÐnot apprising patients of the nature of the TrackCapTM focusing on lung function and safety at clinic visits; and not requiring use of diary cards during the treatment periodÐwere considered to be both necessary to approximate real-world compliance and within the scope of acceptable clinical practice. The value of this study will be re¯ected in future double-blind, crossover studies that compare electronically assessed compliance among dierent forms of asthma treatments and that do not emphasize the compliance objective to patients. Despite the proven bene®ts of inhaled steroids, patients continue, for numerous reasons, to use them inappropriately and inconsistently, thus not achieving the true bene®t (20,26). That patients with asthma prefer oral treatment like za®rlukast to inhaled treatment like beclomethasone dipropionate (27) is not surprising given reported compliance rates. Importantly, such preferences are likely to aect how physicians and other healthcare providers ultimately approach treatment. In conclusion, data from this study show that it is possible to assess adherence and compliance with an oral asthma medication using a medication event monitoring system like TrackCapTM. These results show that compliance with and adherence to a treatment of an oral, twicedaily, maintenance asthma medication, such as za®rlukast, is high.

Acknowledgements The authors thank Suzanne Bristow-Marcalus, BSPh, ELS for writing support, Fiona Campbell for statistical

assistance, Ruth Smithers, Allan A. Harris and James Kennelly PhD for technical assistance, and the clinical investigators for their eort and time: Michael B. Doyle (Belfast, U.K.), Christopher J. Kyle (Belfast, U.K.), J. Earnest Smyth (County Tyrone, U.K.), Mary K. NoelPaton (West Sussex, U.K.), and Constantinos I. Dellaportas (London, U.K.).

REFERENCES 1. Podell RN, Gary LR. Compliance: A problem in medical management. Am Fam Physician 1976; 13: 74± 80. 2. Rand CS, Wise RA, Nides M, et al. Metered-dose inhaler adherence in a clinical trial. Am Rev Respir Dis 1992; 146: 1559±1564. 3. Spector SL. Is your asthmatic patient really complying? Ann Allergy 1985; 55: 552±556. 4. Braunstein GL, Trinquet G, Harper AE, Compliance Working Group. Compliance with nedocromil sodium and a nedocromil/salbutamol combination. Eur Respir J 1996; 9: 893±898. 5. Bosley CM, Parry DT, Cochrane GM. Patient compliance with inhaled medication: Does combining beta agonists with corticosteroids improve compliance? Eur Respir J 1994; 7: 504±509. 6. Mawhinney H, Spector SL, Kinsman RA, et al. Compliance in clinical trials of two non-bronchodilator, antiasthma medications. Ann Allergy 1991; 66: 294±299. 7. Creer TL. Medication compliance and childhood asthma. In: Krasneger NA, Epstein L, Johnson SB, Yae SJ, eds. Developmental Aspects of Health Compliance Behavior. Hillsdale, NJ: Lawrence Erlbaum Associates Publishers, 1993: 303±333. 8. Horn CR, Clark TJH, Cochrane GM. Compliance with inhaled therapy and morbidity from asthma. Respir Med 1990; 84: 67±70. 9. Birkhead G, Attaway NJ, Strunk RC, Townsend MC, Teutsch S. Investigation of a cluster of deaths of adolescents from asthma: Evidence implicating inadequate treatment and poor patient adherence with medications. J Allergy Clin Immunol 1989; 84: 484±491. 10. British Thoracic Association. Death from asthma in two regions of England. Br Med J 1982; 285: 1251± 1255. 11. National Heart, Lung, and Blood Institute. Data Fact Sheet: Asthma Statistics. Bethesda, MD: National Institutes of Health, Public Health Service, 1989. 12. Kelloway JS, Wyatt RA, Adlis SA. Comparison of patients' compliance with prescribed oral and inhaled asthma medications. Arch Intern Med 1994; 154: 1349± 1352. 13. Khunti K. Issues in adolescent asthma: Compliance. Matern Child Health 1996; Feb: 36±38. 14. Rand CS, Wise RA. Measuring adherence to asthma medication regimens. Am J Respir Crit Care Med 1994; 149: 569±576.

858


15. Fabbri LM, Piattella M, Caramori G, Ciaccia A. Oral vs inhaled asthma therapy. Drugs 1996; 52 (Suppl 6):20±28. 16. Spector SL, Smith LJ, Glass M, AccolateTM Asthma Trialists Group. Eects of 6 weeks of therapy with oral doses of ICI 204,219, a leukotriene D4 receptor antagonist, in subjects with bronchial asthma. Am J Respir Crit Care Med 1994; 150: 618±623. 17. Fish JE, Kemp JP, Lockey RF, Glass M, Hanby LA, Bonuccelli CM, Za®rlukast Trialists Group. Za®rlukast for symptomatic mild-to-moderate asthma: a 13week multicenter study. Clin Ther 1997; 19: 675±690. 18. Nathan RA, Bernstein JA, Bielory L, et al. Za®rlukast improves asthma symptoms and quality of life in patients with moderate air¯ow obstruction. J Allergy Clin Immunol 1998;102: 935±942. 19. Grossman J, Smith LJ, Wilson AM, Thyrum PT. Longterm safety and ecacy of za®rlukast in the treatment of asthma: Interim results of an open-label extension trial. Ann Allergy Asthma Immunol 1999; 82: 361±369. 20. Chmelik F, Doughty A. Objective measurements of compliance in asthma treatment. Ann Allergy 1994; 73: 527±532. 21. Spector SL, Kinsman R, Mawhinney H, et al. Compliance of patients with asthma with an experi-

22. 23.

24.

25.

26. 27.

mental aerosolized medication: implications for controlled clinical trials. J Allergy Clin Immunol 1986; 77: 65±70. Cramer JA, Scheyer RD, Mattson, RH. Compliance declines between clinic visits. Arch Intern Med 1990; 150:1509-1510. Simmons MS, Nides MA, Rand CS, Wise RA, Tashkin DP. Trends in compliance with bronchodilator inhaler use between follow-up visits in a clinical trial. Chest 1996; 109: 963±968. Eisen SA, Miller DK, Woodward RS, Spitznagel E, Przybeck TR. The eect of prescribed daily dose frequency on patient medication compliance. Arch Intern Med 1990; 150: 1881±1884. Cramer JA, Mattson RH, Prevey ML, Scheyer RD, Ouellette VL. How often is medication taken as prescribed? A novel assessment technique. JAMA 1989; 261: 3273±3277. Yeung M, O'Connor SA, Parry DT, Cochrane GM. Compliance with prescribed drug therapy in asthma. Respir Med 1994; 88: 31±35. Weinberg EG, Naya I. Treatment preferences of adolescent patients with asthma. Pediatr Allergy Immunol 2000;11: 49±55.