Abstract
Rationale: With the advent of new and expensive therapies for severe refractory asthma, targeting the appropriate patients is important. An important issue is identifying nonadherence with current therapies. The extent of nonadherence in a population with difficult asthma has not been previously reported.
Objectives: To examine the prevalence of nonadherence to corticosteroid medication in a population with difficult asthma referred to a Specialist Clinic and to examine the relationship of poor adherence to asthma outcome.
Methods: General practitioner prescription refill records for the previous 6 months for inhaled combination therapy and short-acting β-agonists were compared with initial prescriptions and expressed as a percentage. Blood plasma prednisolone and cortisol assay levels were used to examine the utility of these measures in assessing adherence to oral prednisolone. Patient demographics, hospital admissions, lung function, oral prednisolone courses, and quality of life data were analyzed to indentify the variables associated with reduced medication adherence.
Measurements and Main Results: A total of 182 patients were assessed. Sixty-three patients (35%) filled 50% or fewer inhaled medication prescriptions; 88% admitted poor adherence with inhaled therapy after initial denial. Twenty-one percent of patients filled more than 100% of presciptions, and 45% of subjects filled between 51 and 100% of prescriptions. Twenty-three of 51 patients (45%) prescribed oral steroids were found to be nonadherent.
Conclusions: A significant proportion of patients with difficult-to-control asthma remained nonadherent to corticosteroid therapy. Objective surrogate and direct measures of adherence should be performed as part of a difficult asthma assessment and are important before prescibing expensive novel biological therapies.
Nonadherence with medication in subjects with different severities of asthma, including subjects with mild and moderate disease has been consistently demonstrated to be common with concomitant asthma morbidity. Data on the degree of nonadherence in a difficult asthma population (symptomatic despite treatment GINA step 4/5) assessed in a specialist service has not been previously presented.
A significant proportion of patients with difficult-to-control asthma are nonadherent to inhaled and oral corticosteroid therapy. This study supports the importance of using objective, surrogate, and direct measures of adherence. Identifying non-adherence in this population is crucially important, given currently available (and imminent) expensive biological therapies but is also central to research efforts to define mechanisms and phenotypes of refractory asthma.
Despite treatment with high-dose therapy, approximately 5% of adult patients remain difficult to control, with persisting symptoms and frequent exacerbations (12). This subgroup of patients is a cause for concern because of the potential consequences of uncontrolled disease, including fatal or near fatal asthma (8, 13–15). Systematic evaluation of subjects with difficult-to-control asthma identifies reasons for persisting symptoms, which are unrelated to disease severity (9, 16). Nonadherence with treatment is likely to be a reason for poor control in this group. A retrospective case control survey identified poor adherence in 22% in 57 children with difficult-to-control asthma (17). The prevalence of poor adherence in adults with difficult asthma (persistent symptoms despite prescribed treatment at GINA step 4/5) (18) remains unclear, and a recent evidence review found little information available in this group (19). Furthermore, the ATS definition of refractory asthma includes the statement “subjects are felt to be generally adherent with therapy” (20), but assessing adherence with asthma therapy can be difficult.
Various methods to measure medication adherence have been studied, but none can be considered the “gold standard.” Direct or objective measures, such as drug assays, pill counts, or electronic monitoring, may be more reliable but have problems with regard to accuracy, cost, and convenience and therefore are not always practical in the clinical setting (4, 21–23). Surrogate measures (e.g., pharmacy refill records) have been found to be a useful and convenient method of measuring and identifying poor adherence and are unobtrusive and inexpensive (24–30).
In patients on oral prednisolone (OP) or high-dose inhaled steroids, it has been suggested that the absence of cortisol suppression in the presence of poor asthma control may potentially identify poor adherence in this group (31). One study identified that approximately 50% of subjects on maintenance oral steroids undergoing detailed inpatient assessment for difficult to control asthma were nonadherent to oral corticosteroids (9).
We hypothesized that nonadherence to medication would be a significant contributing factor in subjects referred to a specialist clinic with difficult-to-control asthma and examined the prevalence of nonadherence to combination inhaled combination therapy (ICT) and OP in sequential patients referred to a Regional Difficult Asthma Service. Some of the results of this study have been previously reported in the form of an abstract (32).
The data presented are a retrospective cross-sectional analysis of adherence and clinical data performed as part of an audit of outcome for the Northern Ireland Regional Difficult Asthma Service. At the time of data collection, there were 188 subjects attending the Service; data were obtained on 182 subjects. The Service assesses approximately 70 new patients a year, and of these approximately 40% are tertiary referrals from other respiratory specialists. The Service has a high discharge rate back to primary care or to the referring respiratory specialist after assessment through the multidisciplinary assessment protocol (16). None of the subjects had nonadherence suspected as a major clinical issue at the time of referral, and all subjects denied nonadherence at the time of first clinical assessment at the clinic. Difficult asthma was defined as persistent symptoms despite treatment at GINA step 4/5 (18).
In Northern Ireland, all prescribed medication is obtained via prescription from a single general practitioner. There is no hospital pharmacy prescribing for outpatients, and, when admitted acutely to a hospital, patients are generally advised to subsequently bring their maintenance medication to the hospital; thus, all prescribed medication is obtained via prescription from a single source, facilitating prescription refill information as a measure of prescribed drug and a surrogate measure of adherence. Single-device ICT (inhaled steroids/long acting β2-agonist) is almost exclusively prescribed in Northern Ireland and had been prescribed in all subjects in this study. General practitioner prescription records were obtained for the previous 6 months for prescription refill rates and were compared with prescribed medication, taking into consideration the number of doses per inhaler and the daily doses prescribed. This was expressed as a percentage of prescribed medication. If a dose or inhaler had been changed during the 6-month period, the rate was adjusted accordingly.
Blood plasma prednisolone and cortisol assay levels were used to examine the potential utility in identifying nonadherence to OP when patients were taking a prescribed short course of rescue prednisolone or were on maintenance prednisolone. All blood samples were taken between 2 and 4 hours after reported ingestion of prednisolone, and in all cases the time and dosage of prednisolone were recorded. Nonadherence to OP was defined as undetectable blood plasma prednisolone with detectable plasma cortisol. We determined the utility of our methods in detecting poor adherence by discussing nonadherence with patients.
Patient demographics, hospital admissions, lung function, OP courses, and QoL data for between-group analyses were obtained retrospectively from clinic notes.
QoL scores were measured using a generic QoL instrument, the EuroQol EQ-5D (33) and the disease-specific Asthma Quality of Life Questionnaire (34). Anxiety and depression were measured using the Hospital Anxiety and Depression Scale (35).
Data were entered into a database, and statistical analysis was performed using SPSS for Windows, version 13 (SPSS, Chicago, IL). Demographic data are presented as mean ± SD or as absolute values. For inhaled medication, we anticipated a spectrum of prescription filling; however, to facilitate the identification of factors associated with poor adherence, we compared data for groups filling prescriptions for greater than 50% of prescribed ICT with data from groups filling 50% or fewer of prescribed doses as previously described (10, 36). Group differences were examined using an independent t test for continuous variables, Chi-square analysis for dichotomous variables, and, where appropriate, Chi-square for trend and Fisher's exact-test. We also analyzed % adherence as a continuous variable using multivariate stepwise linear regression, as previously described (10, 36–38) with log % adherence as the dependent variable to identify variables associated with poor adherence to ICT. A P value of less than 0.05 was considered significant.
One hundred eighty-two consecutive referrals to the difficult asthma clinic were assessed. Of these, 63 patients (35%) filled 50% or fewer prescriptions for ICT, 21% of patients filled more than 100%, and 45% of subjects filled between 51 and 100% of prescribed medication. Prescription filling by quartile is shown in Table 1.
Prescriptions Filled in 6 Mo | Number of Patients |
|---|---|
| 0–25% | 16 (9%) |
| 26–50% | 47 (26%) |
| 51–75% | 34 (19%) |
| 76–100% | 47 (26%) |
| >100% | 38 (21%) |
Demographic details and medication and health care utilization data for the patient population are shown in Table 2, which also shows the comparison between patients filling 50% or fewer of prescriptions for ICT and those filling more than 50% of prescriptions for ICT. One hundred thirteen (62%) of the 182 patients were female, which reflects the usual sex difference in our difficult asthma service. Women were significantly more likely to be nonadherent with ICT than men (42% female; 23% male). There were no age differences between the groups. Patients filling 50% or fewer of prescriptions for ICT were more likely to have been admitted to the hospital on three or more occasions in the previous 12 months.
Prescription Filling | |||||
|---|---|---|---|---|---|
| Demographic and Clinical Data | All Patients (N = 182) | ≤50% ICT (n = 63) | >50% ICT (n = 119) | P Value | |
| Age, yr | 42.1 ± 14.5 | 41.7 ± 13.8 | 42.2 ± 14.8 | 0.84 | |
| Sex, M/F | 69/113 | 16/47 | 53/66 | 0.02 | |
| Source of referral, n (%) | |||||
| Pulmonologist | 67 (37%) | 24 (37%) | 43 (36%) | 1.00 | |
| Primary care | 115 (63%) | 40 (63%) | 75 (64%) | ||
| Smoking history, n (%) | |||||
| Current | 25 (14%) | 8 (13%) | 17 (15%) | 0.55 | |
| Ex | 20 (11%) | 5 (8%) | 15 (13%) | ||
| BMI | 29.1 ± 5.9 | 29.7 ± 5.9 | 28.9 ± 5.9 | 0.43 | |
| Atopy, n (%) | 105 (58%) | 30 (48%) | 75 (63%) | 0.06 | |
| FEV1, % predicted | 72.0 ± 23.9 | 68.7 ± 24.8 | 73.7 ± 23.4 | 0.18 | |
| Admissions in the previous 12 mo, % per number of admissions | 15% = 3 | 25% = 3 | 10% = 3 | 0.02 | |
| 7% = 2 | 5% = 2 | 9% = 2 | |||
| 17% = 1 | 18% = 1 | 16% = 1 | |||
| 61% = 0 | 52% = 0 | 65%= 0 | |||
| Unscheduled visits with GP | 2.0 ± 1.9 | 2.1 ± 1.8 | 1.9 ± 1.8 | 0.58 | |
| Rescue oral steroid courses in the preceding 12 mo | 4.1 ± 3.5 | 4.3 ± 3.4 | 4.0 ± 3.6 | 0.60 | |
| Subjects prescribed maintenance daily oral steroids* | 34/176 (19%) | 14/61 (22%) | 20/115 (17%) | 0.49 | |
| Maintenance daily oral steroid dose, mg | 15.6 ± 11.1 | 18.3 ± 13.1 | 13.7 ± 9.3 | 0.26 | |
| Daily prescribed inhaled steroid dose, μg† | 1,388 ± 550 | 1,494 ± 539 | 1,332 ± 550 | 0.06 | |
| Total short-acting β2-agonist doses in the preceding 6 mo | 1,810 ± 2,181 | 2,013 ± 2,237 | 1,662 ± 2,141 | 0.34 | |
| Owns a compressor for nebulized drugs | 66 (36%) | 31 (49%) | 35 (29%) | 0.01 | |
| Total short-acting β2-agonist nebules prescribed in the preceding 6 mo | 61 ± 133 | 99 ± 182 | 42 ± 87 | 0.03 | |
There was a trend toward higher prescribed doses of daily inhaled steroid and higher total β-agonist inhaler doses used in the 6-month period in the nonadherent group, although this failed to reach statistical significance. Nonadherent patients were more likely to own a compressor and used significantly more nebulized β-agonist over the 6-month period.
Hospital Anxiety and Depression Scale and QoL scores for the entire population are summarized in Table 3, which also shows a comparison between patients filling 50% or fewer of prescriptions for ICT and those filling more than 50% of prescriptions for ICT. Psychiatric “caseness” (i.e., anxiety or depression score ≥11) was seen in 56 patients (31%) for anxiety and in 29 patients (16%) for depression. There were no statistical differences between groups for anxiety and depression scores.
Quality of Life Questionnaire Domains | Prescription Filling | ||||
|---|---|---|---|---|---|
| All Patients (n = 182) | ≤50% ICT (n = 63) | >50% ICT (n = 119) | P Value | ||
| Anxiety | 8.93 ± 4.72 | 9.47 ± 4.70 | 8.66 ± 4.73 | 0.30 | |
| Depression | 6.64 ± 3.93 | 7.04 ± 4.06 | 6.44 ± 3.87 | 0.36 | |
| EURO QoL thermometer | 0.59 ± 0.20 | 0.54 ± 0.20 | 0.62 ± 0.20 | 0.02 | |
| EURO QoL score | 0.54 ± 0.35 | 0.48 ± 0.36 | 0.57 ± 0.34 | 0.12 | |
| AQLQ symptom score | 3.29 ± 1.45 | 2.83 ± 1.04 | 3.52 ± 1.58 | 0.001 | |
| AQLQ activity score | 3.55 ± 1.31 | 3.20 ± 1.13 | 3.72 ± 1.36 | 0.02 | |
| AQLQ emotional score | 3.23 ± 1.58 | 3.01 ± 1.51 | 3.33 ± 1.61 | 0.23 | |
| AQLQ environment score | 3.89 ± 1.49 | 3.70 ± 1.36 | 3.99 ± 1.54 | 0.25 | |
| AQLQ total score | 3.42 ± 1.30 | 3.07 ± 1.03 | 3.59 ± 1.39 | 0.007 | |
Subjects filling 50% or fewer of prescriptions for ICT scored significantly lower in asthma-specific QoL scores for symptoms, activity, and overall score. The mean difference was greater than the minimally clinical important difference of 0.5 (39). Generic QoL scores were also significantly lower using the EQ-5D visual analog scale scores (Table 3).
Linear multivariate stepwise regression analysis with % adherence with ICT as a continuous variable (Table 4) demonstrated three variables to be significantly related to low adherence: female sex (P = 0.001), EURO QoL score (P = 0.02), and hospital admission in the preceding 12 months (P = 0.02).
Unstandardized Coefficients | ||||||
|---|---|---|---|---|---|---|
| Model | B | SE | t Value | P Value | ||
| Constant | 1.614 | 0.070 | 22.961 | 0.000 | ||
| 1 = female; 2 = male | 0.137 | 0.041 | 3.360 | 0.001 | ||
| EURO QoL score | 0.143 | 0.058 | 2.465 | 0.015 | ||
| Admissions in the previous 12 mo | −0.036 | 0.015 | −2.304 | 0.023 | ||
Blood plasma prednisolone levels and cortisol levels were obtained on 51 patients reporting current maintenance OP use (34 patients on maintenance steroids and 17 patients on a short rescue course) and having taken their morning dose in the preceding 2 to 4 hours before their clinic assessment (Table 5). Twenty-three patients (45%) were identified as nonadherent to OP using the criteria defined above. In two patients taking maintenance oral steroids, prednisolone was detectable (316 ± 184 μg/L) with concomitant detectable cortisol (119 ± 2.5 nmol/L) (data not shown). When adherence with prednisolone was discussed, both patients admitted intermittent use of maintenance oral steroids.
Nonadherent (n = 23) | Adherent (n = 26) | |
|---|---|---|
| Blood plasma prednisolone, μg/L | ND | 194 ± 160* |
| Blood plasma cortisol, nmol/L | 212 ± 116 | ND |
Of the 26 subjects who were adherent to prednisolone, 8 (31%) were filling 50% or fewer prescriptions for ICT. Of the 23 subjects who were nonadherent to prednisolone, 35% were adherent to ICS.
Of the 63 patients filling prescriptions for 50% or fewer of the prescribed dose, nonadherence could not be discussed with 12 because they were discharged to their referring specialist or failed to reattend the clinic when offered subsequent review. When combination inhaler/nonadherence findings were discussed with the remaining patients, 88% (45/51) admitted variability in taking ICT at a level consistent with the measured nonadherence based on prescription records. Of the six patients who denied nonadherence with ICT despite having a low prescription filling rate, three had undetectable levels of prednisolone/detectable cortisol or theophylline despite also reporting taking these therapies.
When nonadherence to OP was discussed with patients, 86% (22/25) admitted variability in taking OP; this included the two patients who had detectable prednisolone but also detectable cortisol. Of the three patients who denied nonadherence to OP, two were nonadherent to ICT and denied being nonadherent with these therapies. In the 23 subjects who were nonadherent to prednisolone, 15 (65%) were nonadherent to ICT.
This study demonstrates that a significant proportion of patients with difficult asthma are poorly adherent to inhaled and oral corticosteroid therapy. The issue of nonadherence in asthma is well recognized; however, data in this specific patient group (“difficult asthmatics,” i.e., patients with persistent symptoms despite treatment being prescribed at GINA Step 4/5) have not been previously presented. Defining the scale and identifying nonadherence in this population is important given currently available and other imminent expensive biological therapies, but it is also central to research efforts to define mechanisms and phenotypes of refractory asthma and performing clinical trials in this population.
The inclusion of a milder/less symptomatic asthma control group control would have allowed a more robust examination of adherence and between-group demographic comparisons. However, we did not have access to such a population in the hospital setting, and using a primary care population would have introduced another significant population difference because the lack of hospital/specialist input would be a significant confounding factor. We would anticipate that among patients with milder asthma, nonadherence would also be prevalent, as suggested by previous literature (6, 7). However, we believe the key message of this study—that nonadherence is a significant issue in an unselected population with difficult-to-control asthma and significant asthma-related morbidity referred to a Specialist Regional Centre for assessment (40% tertiary referrals from other pulmonologists)—is valid and needs to be proactively identified and addressed.
Of the 66% of patients filling prescriptions for greater than 50% of prescribed medication, 21% were filling prescriptions for more medication than their fixed dosing prescription would suggest. The issue of “overdosing,” another form of nonadherence, was not a focus of this study, but the use of excessive high-dose steroid medication in this group warrants further study.
All subjects had initially denied poor medication adherence, and poor adherence only became apparent using a combination of surrogate and objective measures that facilitated a discussion with the patient. General practitioner prescription refill records are a surrogate method of measuring adherence, and, as with any surrogate measure, there are potential difficulties. With prescription filling, although patients may be filling prescriptions, they may not be taking the medication and thus are nonadherent. Consequently, our surrogate measure may overestimate adherence in this group. However, patients not filling prescriptions could not be taking more than was assessed because they had not collected a sufficient number of prescriptions. It follows that our overall data may underestimate the scale of the problem in this difficult asthma population but cannot overestimate the prevalence of nonadherence in this group. However, we believe the surrogate measure we have used is valuable because when we discussed nonadherence with patients, the majority (88%) admitted they were poorly adherent with their medication, consistent with their prescription data findings.
There are advantages to using prescription refill data because we were using a “real life” measure of adherence. For example, assessing prescription records retrospectively meant that the study was free from the Hawthorne effect (i.e., improvement of performance when the subject is under observation). Patients who agree to participate in research are more likely than nonparticipants to be adherent with their regimen, and when they are aware that adherence is being measured, patients are more likely to adhere (40). Other advantages of using prescription filling include convenience of use, readily available information that can be obtained before clinic attendance, and minimal cost when compared with other methods (24–29).
Blood plasma prednisolone and cortisol levels have been used previously to examine adherence in difficult asthma, and using similar definitions, one study found that approximately 50% of adult patients with confirmed difficult asthma who were prescribed at least 15 mg/d of prednisolone were nonadherent (9). Prednisolone causes dose-dependent cortisol suppression in normal subjects and in patients with asthma; however, there is variability in individual patient responses (41–43). In subjects with detectable prednisolone and no detectable cortisol, we believe it is reasonable to infer that they are taking prednisolone. In subjects with no detectable prednisolone and normal cortisol or detectable prednisolone and measurable cortisol, interpretation is more complex because the pharmacokinetics of prednisolone have not been precisely studied in this group. Only two subjects had prednisolone detectable with concomitant detectable cortisol, and when the issue was addressed, both subjects admitted intermittent use of daily prescribed maintenance oral steroids. Based on our subsequent patient concordance discussion data, we believe this technique and the definitions that we have applied are useful and should alert the clinician to poor adherence being the probable mechanism for poor asthma control and hence facilitate discussion of adherence with the patient. A prospective pharmacokinetic study of prednisolone and other systemic steroids would be of value in this population.
There are many potential reasons why patients in general may not be adherent to steroid therapy. In a previous qualitative study, we identified that much of the reluctance to take steroid therapy centered on systemic steroids with their greater and more readily identifiable and attributable side-effects (44). However, in this study, subjects who were nonadherent with prednisolone were also more likely to be nonadherent with ICT, suggesting that the reason for nonadherence may apply to all steroid therapies.
Some interesting associations were identified in subjects who were poorly adherent with ICT compared with those who were more adherent with medication. It is significant that the associations identified using regression analysis were similar to those identified when the groups were dichotomized around 50% adherence, supporting the latter analysis, which has been identified as an outcome in previous adherence studies (6, 10, 36, 45).
The three variables associated with poor adherence in the linear regression model were female sex, generic Quality of Life (EuroQoL), and hospital admission in the preceding 12 months. This gender difference has been identified previously (36), but the reason is unknown; nonetheless, it is an area of interest that requires further research.
The association with hospital admission is perhaps not surprising. In subjects with 50% or less adherence, 25% had three or more admissions in the previous year, compared with only 10% of subjects filling more than 50% of precriptions (an adherent group with well-characterized refractory asthma). It is unclear why this level of morbidity does not alter adherence behavior, and the precise mechanism for this poor adherence remains unclear, but the level of recurrent admission suggests an opportunity to identify and address this problem. It also suggests that multiple admissions should raise the suspicion of poor medication adherence.
Generic QoL is associated with poor adherence; however, it is unclear whether this precedes and causes poor adherence or whether poor asthma control and significant morbidity reduces QoL and drives adherence behavior, and again, is an area which requires further study in this group.
Other markers of asthma morbidity (asthma-related QoL, symptom scores, and nebulized short-acting β2–agonist usage) were significantly worse in patients filling 50% or fewer ICT prescriptions compared with more adherent subjects.
These findings indicate that nonadherence is a significant problem in an unselected group of patients attending a Difficult Asthma Service, and one could speculate that if they took regular preventative therapy (as prescribed) their asthma would probably improve substantially. As seen in previous studies, a subjective physician assessment and patient self-report are not good at identifying this problem, and our results endorse the importance of including an objective assessment of adherence as part of a systematic evaluation protocol (9, 16). In addition to identifying subjects who are nonadherent, it is important to recognize and address individual reasons for nonadherence. Our experience suggests that not all patients have the same reasons for poor adherence, and therefore a one-size-fits-all intervention is unlikely to be appropriate. It follows that interventions to improve nonadherence should be individualized to target each patient's particular reason(s).
| 1. | Masoli M, Fabian D, Holt S, Beasley R, Global Initiative for Asthma (GINA) Program. The global burden of asthma: executive summary of the GINA Dissemination Committee report. Allergy 2004;59:469–478. |
| 2. | World Health Organization. The World Health report 2002–reducing risks, promoting healthy life. Geneva, Switzerland: World Health Organization; 2002. |
| 3. | Farmer KC. Methods for measuring and monitoring medication regimen adherence in clinical trials and clinical practice. Clin Ther 1999;21:1074–1090. |
| 4. | Bender B, Rankin A, Tran ZV, Wamboldt FS. Brief-interval telephone surveys of medication adherence and asthma symptoms in the Childhood Asthma Management Program Continuation Study. Ann Allergy Asthma Immunol 2008;101:382–386. |
| 5. | Rand CS, Wise RA. Measuring adherence to asthma medication regimens. Am J Respir Crit Care Med 1994;149:S69–S76. |
| 6. | Cochrane GM, Mohan BV, Downs KE, Mauskopf J, Rami H. Inhaled corticosteroids for asthma therapy: patient compliance, devices and inhalation technique. Chest 2000;117:542–550. |
| 7. | Cereveri I, Locatelli F, Zoia MC, Corsico A, Accordini S, de Marco R; on behalf of the ECRHS. International variations in asthma treatment compliance: the results of the European Community Respiratory Health Survey (ECRHS). Eur Respir J 1999;14:288–294. |
| 8. | Chung KF, Godard P, Adelroth E, Ayres J, Barnes N, Barnes P, Bel E, Burney P, Chanez P, Connett G, et al. Difficult/therapy-resistant asthma: the need for an integrated approach to define clinical phenotypes, evaluate risk factors, understand pathophysiology and find novel therapies. ERS Task Force on Difficult/Therapy-Resistant Asthma. European Respiratory Society. Eur Respir J 1999;13:1198–1208. |
| 9. | Robinson DR, Campbell DA, Durham SR, Pfefer J, Barnes PJ, Chung KF. Systematic assessment of difficult to treat asthma. Eur Respir J 2003;22:478–483. |
| 10. | Krishnan JA, Riekert KA, McCoy JV, Stewart DY, Schmidt S, Chanmugam A, Hill P, Rand CS. Corticosteroid use after hospital discharge among high-risk adults with asthma. Am J Respir Crit Care Med 2004;170:1281–1285. |
| 11. | Williams LK, Pladevall M, Xi H, Peterson E, Joseph C, Lafata JE, Ownby DR, Johnston CC. Relationship between adherence to inhaled corticosteroids and poor outcomes among adults with asthma. J Allergy Clin Immunol 2004;114:1288–1293. |
| 12. | Barnes PJ, Woolcock AJ. Difficult asthma. Eur Respir J 1998;12:1209–1218. |
| 13. | Innes NJ, Reid A, Halstead J, Watkin SW, Harrison BD. Psychosocial risk factors in near-fatal asthma and in asthma deaths. J R Coll Physicians Lond 1998;32:430–434. |
| 14. | Wareham NJ, Harrison BD, Jenkins PF, Nicholls J, Stableforth DE. A district confidential enquiry into deaths due to asthma. Thorax 1993;48:1117–1120. |
| 15. | Mohan G, Harrison BD, Badminton RM, Mildenhall S, Wareham NJ. A confidential enquiry into deaths caused by asthma in an English health region: implications for general practice. Br J Gen Pract 1996;46:529–532. |
| 16. | Heaney LG, Conway E, Kelly C, Johnston BT, English C, Stevenson M, Gamble J. Predictors of therapy resistant asthma: outcome of a systematic evaluation protocol. Thorax 2003;58:561–566. |
| 17. | Ranganathan SC, Payne DN, Jaffe A, McKenzie SA. Difficult asthma: defining the problem. Pediatr Pulmonol 2001;31:114–120. |
| 18. | Global Initiative for Asthma (GINA). Global strategy for asthma management and prevention. 2008; Available from: http://www.ginasthma.org/guideline (accessed January 12, 2009). |
| 19. | British Thoracic Society Scottish Intercollegiate Guideline Network. British guideline on the management of asthma. Thorax 2008;63(suppl 4): iv, 1–121. |
| 20. | Wenzel SE, Fahy JV, Irvin C, Peters SP, Spector S, Szefler SJ. Proceedings of the ATS workshop on refractory asthma: current understanding, recommendations, and unanswered questions. Am J Respir Crit Care Med 2000;162:2341–2351. |
| 21. | Haynes RB, Taylor DW, Sackett DL, Gibson ES, Bernholz CD, Mukherjee J. Can simple clinical measurements detect patient noncompliance? Hypertension 1980;2:757–764. |
| 22. | Morisky DE, Green LW, Levine DM. Concurrent and predictive validity of a self-reported measure of medication adherence. Med Care 1986;24:67–74. |
| 23. | Cochrane GM, Horne R, Chanez P. Compliance in asthma. Respir Med 1999;93:763–769. |
| 24. | Sherman J, Hutson A, Baumstein A, Hendeles L. Telephoning the patient's pharmacy to assess adherence with asthma medications by measuring refill rate for prescriptions. J Pediatr 2000;136:532–536. |
| 25. | Hamilton RA, Briceland LL. Use of prescription-refill records to assess patient compliance. Am J Hosp Pharm 1992;49:1691–1696. |
| 26. | Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. J Clin Epidemiol 1997;50:105–116. |
| 27. | Choo PW, Rand CS, Inui TS, Lee MT, Cain E, Cordeiro-Breault M, Canning C, Platt R. Validation of patient reports, automated pharmacy records, and pill counts with electronic monitoring of adherence to antihypertensive therapy. Med Care 1999;37:846–857. |
| 28. | Bieszk N, Patel R, Heaberlin A, Wlasuk K, Zarowitz B. Detection of medication non adherence through review of pharmacy claims data. Am J Health Syst Pharm 2003;60:360–366. |
| 29. | Watts RW, McLennan G, Bassham I, El Saadi O. Do patients with asthma fill their prescriptions? a primary compliance study. Aust Fam Physician 1997;26:S4–S6. |
| 30. | Kelloway JS, Wyatt R, Adlis SA. Comparison of patients' compliance with prescribed oral and inhaled medications. Arch Intern Med 1994;154:1349–1352. |
| 31. | AbuRuz S, Heaney LG, Millership J, Gamble J, McElnay J. A cross sectional study evaluating the relationship between cortisol suppression and asthma control in patients with difficult asthma. Br J Clin Pharmacol 2007;63:110–115. |
| 32. | Gamble J, Heaney L. Non-adherence: a major problem in difficult asthma [abstract]. Am J Respir Crit Care Med 2007;175:A686. |
| 33. | The EuroQol Group. Euroqol: a facility for the measurement of health-related quality of life. Health Policy 1990;16:199–208. |
| 34. | Juniper EF, Guyatt GH, Ferrie PJ, Griffith LE. Measuring quality of life in asthma. Am Rev Respir Dis 1993;147:832–838. |
| 35. | Zigmond AS, Snaith PR. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983;67:361–370. |
| 36. | Smith BS, Krishnan JA, Bilderback A, Riekert KA, Rand CS, Bartlett SJ. Depressive symptoms and adherence to asthma therapy after hospital discharge. Chest 2006;130:1034–1038. |
| 37. | Wells K, Pladevall M, Peterson EL, Campbell J, Wang M, Lanfear DE, Williams LK. Race-ethnic differences in factors associated with inhaled steroid adherence among adults with asthma. Am J Respir Crit Care Med 2008;178:1194–1201. |
| 38. | van Dellen QM, Stronks K, Bindels PJ, Ory FG, van Aalderen WM; PEACE Study Group. Adherence to inhaled corticosteroids in children with asthma and their parents. Respir Med 2008;102:755–763. |
| 39. | Juniper EF, Guyatt GH, Willan A, Griffith LE. Determining a minimal important change in a disease-specific quality of life questionnaire. J Clin Epidemiol 1994;47:81–87. |
| 40. | Ramsay JA. Participants in non compliance research: compliant or noncompliant? Med Care 1982;20:615–622. |
| 41. | Adair CG, McCallion O, McElnay JC, Scott MG, Hamilton BA, McCann JP, Stanford CF, Nicholls DP. A pharmacokinetic and pharmacodynamic comparison of plain and enteric-coated prednisolone tablets. Br J Clin Pharmacol 1992;33:495–499. |
| 42. | Wilson AM, Lipworth BJ. Short-term dose-response relationships for the relative systemic effects of oral prednisolone and inhaled fluticasone in asthmatic adults. Br J Clin Pharmacol 1999;48:579–585. |
| 43. | Wilson AM, McFarlane LC, Lipworth BJ. Systemic bioactivity profiles of oral prednisolone and nebulized budesonide in adult asthmatics. Chest 1998;114:1022–1027. |
| 44. | Gamble J, Fitzsimons D, Lynes D, Heaney LG. Difficult asthma: people's perspectives on taking corticosteroid therapy. J Clin Nurs 2007;16:59–67. |
| 45. | Cooper WO, Hickson GB. Corticosteroid prescription filling for children covered by Medicaid following an emergency department visit or a hospitalization for asthma. Arch Pediatr Adolesc Med 2000;115:1111–1115. |
