Medication prescribing for asthma and COPD: a register-based cross-sectional study in Swedish primary care

Asthma and chronic obstructive pulmonary disease (COPD) are respiratory diseases representing a major burden in health care [1, 2]. The prevalence varies between countries, with an estimated prevalence of 8% for asthma, and approximately 4-6% of the adult population having COPD in Sweden [3, 4]. There are a number of national and international guidelines for the diagnosis and management of asthma and COPD [2, 5, 6]. Asthma medication is approved for use in both asthma and in COPD and for COPD there is also specific treatment, such as anticholinergics for inhalation, not approved for use in asthma [6].

The diagnosis and management of asthma and COPD often stray from guidelines [7, 8] leading to both under-diagnosis, misdiagnosis and inappropriate prescribing practices [9‐11]. Particularly for children, asthma medication is often prescribed without diagnosed asthma [12‐15]. One third of all preschool children experience periods of wheezing [14] and it may not be easy to differentiate asthma from other wheezing disorders. In the absence of objective measurement, asthma medication is often used as diagnostic tool [15]. However, for patients from the age of six years; the diagnosis can be made with reasonable certainty based on objective measurements [3]. Thus the discrepancy between prescribed asthma medication and diagnosed asthma can potentially be reduced from the age of six. However, in the Netherlands, Zuidgeest et al. found that the gap also continues into adolescent age and they conclude that this could be due to factors such as under-diagnosis, off-label prescribing and the use of asthma medications as a diagnostic tool [12]. Few studies have explored the discrepancy between prescribed medication and diagnosed asthma or COPD in adults. In the Netherlands, Lucas et al. [16] found that only 74% of the prescriptions of inhaled corticosteroids were prescribed to patients actually having asthma or COPD, indicating a discrepancy also between prescribed medication and COPD. The use of prescribed medication as a tool in the diagnostic assessment was shown in a study from Germany [8]. In this study, trial of medication was used in 22% of the patients with asthma, but among the patients with COPD, the number was only 5%. Trial of medication thus appears to exist in both asthma and COPD, but to a different extent.

Off-label prescribing is common in many diseases [13, 17]. In such prescribing, a registered medicine is used in a manner not comprised in the license, including the use of drugs in non-approved age-groups (i.e. children), unrecommended doses, formulations or indications [18]. In the US, 40% of the medications for asthma are prescribed off-label [17]. Even though not recommended, for some diagnoses it may sometimes be clinically appropriate and might reflect innovative clinical practice. However, in a US study, 73% of the off-label prescriptions had no or little scientific support [18], and were associated with a number of safety, clinical and ethical issues. Published reports thus indicate that there is a gap between prescribed asthma medication and physician-diagnosed asthma among children and adolescents. However, only few studies have explored this discrepancy among adults, where asthma medication is also indicated for the treatment of COPD.

Characteristics of the population in Skaraborg, the selected population from the PHCs and the random sample is presented in Table 1.

During 2004-2005 medication indicated for asthma and COPD were prescribed to 5.6% (Group A; n = 14 101) of the individuals in Skaraborg primary care. Within this population of medication users, an asthma diagnosis was recorded for 5876 individuals (42%), 1116 (8%) were diagnosed with COPD and 545 (4%) with both COPD and asthma. The remaining 6564 individuals (46%) were lacking a recorded ICD-code for any of the diagnoses.

The results from Group B demonstrated that an asthma diagnosis was recorded for totally 9296 individuals, 2288 had COPD and 744 had both asthma and COPD. Prescription of medication were lacking among 3420 of the individuals with asthma (37%), 1172 with COPD (51%) and among 199 with both asthma and COPD (27%). The prevalence of asthma and COPD based solely on recorded diagnosis was for the Skaraborg population 3.6% and 1.2% respectively. The lack of congruence between medication users and individuals with a recorded diagnosis of asthma or COPD is presented graphically in a Venn diagram in Figure 2.

The prevalence of medication in relation to the prevalence of recorded diagnoses, in different age-groups is presented in Figure 3. Lower rate of diagnoses than prescriptions was apparent in all age-groups. Prevalence of medication use was lowest among the youngest patients and increased with age. The average discrepancy (calculated as a weighted mean) between medication prescribing and diagnosis was 2.5 percentage points.

This is an extensive survey of drug prescribing among a large broadly representative group of patients and based on individual patient records. Our study indicate that medication indicated for asthma and COPD seems to be both over- and under-used; over-used when only half of the medication users actually had one of the recommended diagnoses of asthma or COPD recorded. Under-used when individuals having asthma or COPD were lacking prescriptions of medication. We also show that the use of asthma medication as a diagnostic tool and off-label prescribing were common. Our results are of interest when auditing prescribing in general practice. It should also be observed by researchers in the field studying the epidemiology of asthma and COPD medication use, as discussed below.

The Venn diagram in Figure 2, shows prescribing of medication in relation to diagnoses of asthma or COPD. The figure illustrates the discrepancy between the prevalence of asthma and COPD and the prescribing of medication indicated for these conditions. A considerable proportion of the individuals with asthma or COPD did not have any medication prescribed during the study period, which likely indicates under-medication [9, 21]. It could also be due to previous incorrect diagnoses of asthma or COPD of individuals not having the recorded disease [22, 23]. For COPD it could also be due to exclusive prescribing of anticholinergics (specific only for COPD), without simultaneous prescribing of SABA, LABA or ICS. Approximately half of the individuals prescribed a medication had neither a diagnosis of asthma nor of COPD recorded, which probably indicate over-medication (i.e. off-label prescribing) [13, 17, 24]. It could also be due to under-diagnosis of asthma or COPD [10, 11], which is supported by the low prevalence figures of 3.6% and 1.2%, as compared to the national prevalence of 8% and 4%, respectively [25]. Similar figures have previously been reported from other regions of the Swedish primary care [26].

We have shown that the gap between asthma medication and recorded diagnoses, previously shown mostly in children and adolescents [12, 13], exists in all age-groups (Figure 3). The average prevalence of medication use was 5.6%, in line with previous findings [12, 21]. However, we found that there are fewer children using asthma medication, as compared to the results by Zuidgeest et al. [12] and a decrease in volume of prescribing in the elderly population, which runs counter to the findings of another Swedish study [21]. These differences could be due to inclusion of anticholinergics included in the total volume of drugs in this age-group, in the study by Haupt et al. [21]. It could also be due to hospital doctors’ prescribing medication for both the oldest and the youngest, which are not shown in data from primary care [27], and therefore could be interpreted as an absence of medication.

The detailed examination of patient records in the random sample showed that more than half of the undiagnosed medication users (54%) had the medication prescribed off-label. Many previous studies have been performed concerning off-label prescribing among children [24]. Our study shows that off-label prescribing of asthma medication also is common among adults, and used to treat patients with diagnoses of acute bronchitis and cough. Similar findings were reported for ß₂-adrenergics for acute bronchitis [28]. Even though the Cochrane group found no evidence for such prescribing [29]; the practice exists, suggesting lack of adherence to evidence based prescribing. The use of medication off-label may be a waste of resources in health care. Moreover, asthma medication represents the most frequently reported cause of adverse reactions among children and adolescents, since half of them were prescribed off-label [30]. Therefore the practice also could constitute a risk for these patients.

We found that asthma medication was often used as a part of the diagnostic assessment (30%) and 16% of the individuals, had provisional diagnoses of asthma or COPD, i.e. asthma or COPD based on doctor’s written notes and no ICD-code recorded (Table 3). Low levels of PEF and spirometry are shown in Table 3, similar results are reported in our previous study [7], indicating low adherence to guidelines. Among the individuals where the medication was used as a part of the diagnostic assessment, the trial of medication may have been used for diagnosis rather than peak expiratory flow measurement or spirometry, as also reported elsewhere [8].

Among patients with asthma, 79% were issued a medication once or twice during the two-year period (Table 3). When also considering those with no prescriptions, we conclude that many patients may be inadequately treated, which has also been shown in other studies [9, 10]. For patients with COPD, the situation was different. They had more prescriptions issued (Table 3) and they contributed extensively to the total population of asthma medication users in the older age-groups (Figure 3). Assuming one year’s supply for each prescription (which is common in Sweden) our findings suggest over-medication among patients with COPD, since approximately half of the patients were prescribed three or more medications during the two-year period (Table 3). However, since under-detection of COPD is common [11], the patients detected and diagnosed may be those with more severe COPD and therefore in need of more medication. This interpretation is further supported by the more common use of LABA and fixed combination of LABA/ICS among these patients, and the fact that these medications are to be used in the moderate to severe stages of COPD [6].

Among the 71 patients with COPD, only 37 (52%) had spirometry performed. Related data from USA and Denmark show similar levels [31, 32], which has also been reported in our previous study [7]. This is a clear shortfall in the diagnosis of COPD and a lack of clinical resource and training may be a liable factor in this case. However, access to spirometry equipment was reported by 95% of the primary health care centers in the same region during the time of the study [33] and the low levels of spirometry found in the present study are to be explained by other factors than availability of spirometry. Against this background of deficiencies in objective diagnostic measures, a trial of medication to help diagnose asthma and COPD appears to be widespread practice.

The use of asthma medication as a proxy for disease prevalence is an on-going debate. Our study confirms previous results of incongruenty between prescription and diagnosis. In the Netherlands, using data from primary care, a single prescription of asthma medication could identify 95% of adult asthma patients but at the same time 30% of all asthma medication users with a single prescription were non-asthma patients [34]. On the other hand, several studies have shown a close resemblance between prevalence of medication and prevalence of disease [10, 21]. However, in these studies the use of non-linked data sets and the lack of information about diagnoses in databases with dispensed medication are limitations. Our study indicates that the prevalence of asthma and COPD cannot be reliably estimated from the prevalence of prescribing of medication indicated for either of the conditions.

There was a large discrepancy between prescribing of medication and the prevalence of diagnosed asthma and COPD. Consequently, the prevalence of prescriptions of medications indicated for asthma and COPD should not be used to estimate the prevalence of these conditions. Asthma medication was often prescribed outside recommendations and used both as a diagnostic tool and in an off-label manner. Therefore, the prescribing of medications for asthma and COPD currently does not adhere to national clinical guidelines. More efforts should be made to improve the prescribing of medication indicated for asthma and COPD so that they align with current guidelines. We recommend that the prescribing of these medications and the medical management of both asthma and COPD are regularly audited to improve the prescribing practices in Skaraborg.