Trends and predictors of specialist assessments in oral corticosteroid treated asthma among young adults.

Background: Repeated oral corticosteroid use indicates uncontrolled disease among asthma patients, and referral for asthma specialist assessment is recommended. We aimed to describe trends and predictors associated with specialist contacts among young adults with asthma and repeated oral corticosteroid use.
Methods: Individuals aged 18-45 years with two or more dispensed asthma medication prescriptions and two dispended oral corticosteroid prescriptions (including short-term and long-term treatments) within 12 months during 1999-2018 were identified by use of Danish healthcare registers. The frequency of specialist contacts within 1 year of follow-up was assessed among individuals without previous specialist contacts within 5 years of inclusion. Factors associated with specialist contact were identified using logistic regression models. Furthermore, oral corticosteroid prescriber sources were assessed.
Results: For the 11 223 individuals included, 2444 (22%) had previous specialist-contact care within 5 years prior of inclusion, and additionally 926 (8.3%) within 1 year of follow-up. Among those without previous specialist contacts (n=8779), the frequency of incident specialist contacts within 1 year of follow-up increased from 6.3% in 1999 to 18% in 2017. Factors associated with incident specialist contacts included dispensing ≥12 short-acting β-agonist canisters and previous asthma-related emergency department visits and hospitalisations. The majority of oral corticosteroid prescriptions at baseline (71%) were prescribed by general practitioners, although with decreasing proportions from 1999 to 2018. Conclusions: The majority (70%) of young adults with asthma and repeated oral corticosteroid use do not seem to receive specialist assessment in Denmark. This highlights a potential room for improvement in the patient referral pathway for at-risk asthma patients.

Background

Asthma is a common inflammatory airway disease with an estimated prevalence of 8–10% among adults in Denmark [1, 2]. Most patients with asthma are managed in primary care, with the option of referral for asthma specialist assessment, e.g. in case of uncertain diagnosis or severe or uncontrolled disease. Despite the advances in asthma understanding and management in recent decades, poor asthma control is prevalent in more than one in three patients with severe asthma and one in four with mild–moderate asthma in Scandinavia [3-5], with major consequences for patients’ quality of life, as well as societal costs [5-7].

Oral corticosteroids (OCS) are used for treating uncontrolled asthma, either as short-term courses for severe exacerbations or long-term treatments for severe asthma that remains uncontrolled despite otherwise optimised treatment [8]. Although new therapies for controlling asthma have emerged over the years, OCS continue to be used frequently in asthma management [9] with no reduction in the prevalence of OCS users among young adults with asthma in Denmark during the past two decades [10]. Recently, international experts have proposed that a cumulative OCS exposure of 0.5–1 g per year (equivalent to two to four OCS exacerbation courses) is indicative of poor asthma control [11] and that patients receiving two or more courses within a year should be considered referred for specialist assessment [12]. Similarly, since 2014, the Global Initiative for Asthma (GINA) has recommended referral for expert advice in case of long-term or frequent OCS use, e.g. two or more courses a year [13]. Recent studies have mainly focused on the referral pathways among severe asthma populations [3, 4, 14, 15], but if the overall OCS use in asthma management is to be minimised, a focus on general asthma populations is called for. A great deal of inappropriate OCS use occurs in mild–moderate asthma which may be poorly controlled due to underuse of ICS and/or poor adherence [9, 16]. The most important problem in suboptimally treated asthma is recurrent exacerbations, decline of pulmonary function and OCS-associated side-effects [5, 7, 16]. Growing evidence suggests that receiving even a few OCS courses is associated with long-term side-effects in general asthma populations [9, 17–19], emphasising a need for easy-to-recall red flags for the identification of at-risk patients in broader asthma populations who would benefit from a second opinion from a specialist.

Therefore, we aimed to describe trends and factors associated with specialist assessment in a nationwide cohort of young adults with asthma and repeated oral corticosteroid use over a 20-year period using population-based healthcare registers.

Materials and methods

Design and data sources

We performed a register-based open cohort study with a study period from 1999 to 2018. Data from nationwide administrative and healthcare registers were provided by the Danish Health and Medicines Authority and included data on basic demographics [20], drug prescriptions filled at community pharmacies [21], procedures and diagnoses from hospitals [22] and services from private practices [23]. Pseudonymised data were linked on an individual level using the civil registration number unique to all Danish citizens [24].

Study population

A study population of young adults with asthma and repeated OCS use was based on validated methods and identified as individuals aged 18–45 years with two or more redeemed asthma medication prescriptions (including inhaled corticosteroids (ICS), selective β2-agonists, leukotriene receptor antagonists and xanthines) [25, 26], and two OCS prescriptions within 12 consecutive months [8, 12] (i.e. the baseline period) using the second OCS prescription as index date. OCS prescriptions included prednisolone and prednisone independent of dose and duration. Exclusion criteria included hospital-given diagnoses of COPD or cystic fibrosis, and <5 years of available data prior to cohort entry (i.e. recent migrations, etc.). Furthermore, individuals with comorbidities often treated with OCS (including sarcoidosis, primary adrenocortical insufficiency, pneumonitis, inflammatory bowel disease, inflammatory polyarthropathies, systemic connective tissue disorders, inflammatory spondylopathies and/or malignancy, as defined in supplementary table S1) were excluded at the index date and censored during follow-up upon incident diagnosis. All individuals were followed for a maximum of 5 years after the index date, until death or migration. The patient selection flowchart is shown in figure 1 and the study design is shown in supplementary figure S1.

FIGURE 1

Flowchart of patient selection.

Covariates

Baseline characteristics at index date included sex, age, marital status and region of residency. Asthma medication use, number of asthma-related emergency department visits and hospitalisations, and co-medication use (including systemic antibiotics, systemic antihistamines, nasal corticosteroids, antidepressants, anti-acid drugs, anti-obesity drugs, antidiabetic drugs excluding insulins and bisphosphonates) were assessed during the baseline period. ICS use was categorised as no use, low dose (≤400 μg per day) or medium/high dose (>400 μg per day) in budesonide equivalents [8]. To enable comparison of short-acting β2-agonists (SABA) canisters, one canister was defined as 200 doses (puffs) irrespective of dosage and strength. SABA use was categorised as low use (0–<3 canisters), increased use (3–<12 canisters) and excessive use (≥12 canisters). OCS prescriber source was categorised as general practitioners, private specialists, hospital physicians and others. Years since the first asthma medication dispensing from index date was used as an indicator of number of years lived with asthma.

Specialist assessment

Patient contacts with specialised care were defined by presence of an outpatient hospital contact with a relevant asthma-related diagnosis code as defined by the Danish National Database for Asthma (DrAstma) [27] or as a contact with a private specialist with a relevant pulmonary service code, as described previously [3] (further specified in supplementary table S1).

The main outcomes of interests were the proportion of individuals with a specialist contact during a 5-year period leading up to time of inclusion and the proportion of incident specialist contacts within 1 year of follow-up (i.e. among those without previous specialist contacts).

Baseline characteristics were evaluated for their potential association with incident specialist assessments within 1 year of follow-up among patients included during 2014–2017. The analysis was restricted to 2014–2017 due to GINA first implementing the recommendation of referral for specialist advice if the patient had used repeated OCS (e.g. two courses or more a year) in 2014. Furthermore, this was done in order to increase the clinical relevance of the results.

The time for achieving incident specialist assessment was evaluated within a 5-year follow-up window.

Statistical analyses

Categorical variables were summarised as number and percentage and compared using Chi-squared tests of independence. Continuous variables were reported as median and interquartile range (IQR) and compared by nonparametric equality-of-medians test. Among individuals included during 2014–2017, associations between baseline characteristics (covariates) and receiving specialist assessment within 1 year of follow-up (outcome) were evaluated by multivariable logistic regression and reported as crude and adjusted odds ratios with 95% confidence intervals. The time for first specialist contact within 5 years of follow-up was illustrated graphically.

Two sensitivity analyses were performed with alternative definitions of “repeated OCS users” as patients with three and four OCS prescriptions within the baseline year, respectively, in order to explore the impact of choosing other thresholds for a potential guideline recommendation.

A supplementary post hoc analysis was performed on a subpopulation with possible severe asthma defined by GINA steps 4–5 (use of medium or high-dose ICS plus at least one add-on treatment within the baseline period) [8].

All data were analysed using Stata version 17.0 (StataCorp, College Station, TX, USA).

Results

Baseline characteristics

Baseline characteristics are presented in table 1. A total of 11 223 individuals with asthma (62% female; median (IQR) age 36 (29–41) years) were included in the study population as repeated OCS users, of whom 2444 (22%) had a specialist contact within 5 years prior to inclusion. Patients with previous specialist contacts were younger (35 years versus 37 years, p<0.001) and more often female (67% versus 61%, p<0.001). Furthermore, they were more often treated with medium/high-dose ICS and add-on therapies and less often had excessive use of SABA (table 1).

TABLE 1

Baseline characteristics of young adults with asthma and repeated oral corticosteroid use stratified according to previous specialist contacts (within 5 years of index date)

	All patients	Previous specialist contacts	No previous specialist contacts	p-value
Individuals	11 223	2444	8779
Female	7003 (62.4)	1637 (67.0)	5366 (61.1)	<0.001
Age	36 (29–41)	35 (26–41)	37 (30–42)	<0.001
18–25 years	1761 (15.7)	555 (22.7)	1206 (13.7)	<0.001
26–35 years	3406 (30.3)	734 (30.0)	2672 (30.4)	0.709
36–45 years	6056 (54.0)	1155 (47.3)	4901 (55.8)	<0.001
Marital status
Unmarried	3434 (30.6)	928 (38.0)	2506 (28.5)	<0.001
Married/registered partnership	4717 (42.0)	933 (38.2)	3784 (43.1)	<0.001
Divorced/widowed	1242 (11.1)	223 (9.1)	1019 (11.6)	<0.001
Other/missing	14 (0.1)	5 (0.2)	9 (0.1)	0.203
Region of residency
Capital	3262 (29.1)	888 (36.3)	2374 (27.0)	<0.001
Zealand	1734 (15.5)	236 (9.7)	1498 (17.1)	<0.001
North Denmark	1094 (9.7)	189 (7.7)	905 (10.3)	<0.001
Central Denmark	2595 (23.1)	473 (19.4)	2122 (24.2)	<0.001
Southern Denmark	2520 (22.5)	654 (26.8)	1866 (21.3)	<0.001
Missing	18 (0.2)	n<5
Years since first asthma drug dispensing (any time before index date)	7 (4–13)	9 (4–14)	7 (4–12)	<0.001
Concurrent asthma medication
ICS
No use	1735 (15.5)	200 (8.2)	1535 (17.5)	<0.001
Low dose	5696 (50.8)	1208 (49.4)	4488 (51.1)	0.143
Medium/high dose	3792 (33.8)	1036 (42.4)	2756 (31.4)	<0.001
LABA	6296 (56.1)	1790 (73.2)	4506 (51.3)	<0.001
LTRA	1876 (16.7)	732 (30.0)	1144 (13.0)	<0.001
LAMA	344 (3.1)	126 (5.2)	218 (2.5)	<0.001
SABA canisters
0–<3	4690 (41.8)	978 (40.0)	3712 (42.3)	0.046
3–<12	4749 (42.3)	1145 (46.8)	3604 (41.1)	<0.001
≥12	1784 (15.9)	321 (13.1)	1463 (16.7)	<0.001
Co-medication
Antibiotics	8009 (71.4)	1731 (70.8)	6278 (71.5)	0.511
Antihistamines	3655 (32.6)	1042 (42.6)	2613 (29.8)	<0.001
Nasal corticosteroids	2983 (26.6)	942 (38.5)	2041 (23.2)	<0.001
Antidepressants	1628 (14.5)	299 (12.2)	1329 (15.1)	<0.001
Anti-acid drugs	1813 (16.2)	438 (17.9)	1375 (15.7)	0.008
Anti-obesity drugs	355 (3.2)	57 (2.3)	298 (3.4)	0.007
Antidiabetic drugs, excluding insulins	157 (1.4)	36 (1.5)	121 (1.4)	0.698
Bisphosphonates	16 (0.1)	n<5		0.762
Asthma-related ED visits
1	431 (3.8)	147 (6.0)	284 (3.2)	<0.001
2	73 (0.7)	30 (1.2)	43 (0.5)	<0.001
≥3	47 (0.4)	13 (0.5)	34 (0.4)	0.374
Asthma-related hospitalisations
1	1097 (9.8)	336 (13.7)	761 (8.7)	<0.001
2	276 (2.5)	133 (5.4)	143 (1.6)	<0.001
≥3	141 (1.3)	70 (2.9)	71 (0.8)	<0.001

Data are presented as n, n (%) or median (interquartile range), unless otherwise stated. ICS: inhaled corticosteroid; LABA: long-acting β2-agonist; LTRA: leukotriene receptor antagonist; LAMA: long-acting muscarinic antagonist; SABA: short-acting β2-agonist; ED: emergency department.

Trends in specialist assessment

Among those without previous specialist contacts, 11% (926 out of 8779) had an incident specialist contact within 1 year of follow-up, resulting in 70% of the total cohort (7853 out of 11 223) not meeting the primary end-point of specialist assessment either 5 years prior to or 1 year after inclusion. Annual cross-sectional analyses showed that the proportion of incident specialist contacts within 1 year of follow-up increased from 6.3% in 1999 to 18% in 2017 (figure 2).

FIGURE 2

Frequency of incident specialist assessments within 1 year of follow-up.

Characteristics associated with specialist assessment

Several characteristics appeared to be associated with incident specialist assessment among individuals included during 2014–2017. The strongest associated factors included asthma-related emergency department visits (OR 3.76, 95% CI 2.14–6.61), asthma-related hospitalisations (OR 3.19, 95% CI 2.16–4.71), medium/high dose ICS (OR 1.80, 95% CI 1.16–2.80) and two or more add-on controllers (OR 1.72, 95% CI 1.09–2.71). Patients of higher age (36–45 years), divorced/widowed patients and patients residing outside the Capital region and Zealand were less likely to receive specialist assessment (table 2). However, when adjusting for the other factors in the model, only asthma-related emergency department visits (OR 2.62, 95% 1.42–4.84), hospitalisations (OR 2.59, 95% CI 1.71–3.90), ≥12 SABA canisters (OR 1.78, 95% 1.01–3.14) and residence in North Denmark (OR 0.65, 95% 0.44–0.97) achieved statistically significant p-values <0.05.

TABLE 2

Factors associated with specialist assessment among young adults with asthma and repeated oral corticosteroid use included during 2014–2017 (only individuals without previous specialist contacts)

	Crude OR (95% CI)	p-value	Adjusted# OR (95% CI)	p-value
Female (reference)	1.00		1.00
Male	0.89 (0.67–1.19)	0.443	0.93 (0.69–1.25)	0.638
Age
18–25 years (reference)	1.00		1.00
26–35 years	0.87 (0.58–1.29)	0.477	0.94 (0.61–1.45)	0.773
36–45 years	0.63 (0.44–0.92)	0.015	0.74 (0.47–1.15)	0.181
Marital status
Unmarried (reference)	1.00		1.00
Married/registered partnership	0.81 (0.59–1.09)	0.164	0.95 (0.66–1.37)	0.793
Divorced/widowed	0.55 (0.33–0.93)	0.027	0.57 (0.32–1.02)	0.057
Other/missing	0.83 (0.50–1.40)	0.490	0.79 (0.46–1.38)	0.411
Region of residency
Capital (reference)	1.00		1.00
Zealand	0.73 (0.49–1.08)	0.119	0.86 (0.57–1.29)	0.457
North Denmark	0.61 (0.42–0.88)	0.009	0.65 (0.44–0.97)	0.036
Central Denmark	0.55 (0.32–0.93)	0.027	0.58 (0.33–1.00)	0.052
Southern Denmark	0.61 (0.40–0.94)	0.024	0.72 (0.46–1.14)	0.160
Concurrent asthma medication
ICS
No use (reference)	1.00		1.00
Low dose	1.75 (1.18–2.59)	0.006	1.52 (0.99–2.35)	0.057
Medium/high dose	1.80 (1.16–2.80)	0.009	1.31 (0.78–2.21)	0.303
Add-on controllers (LABA, LAMA, LTRA)
0 (reference)	1.00		1.00
1	1.17 (0.87–1.57)	0.308	1.05 (0.76–1.45)	0.766
≥2	1.72 (1.09–2.71)	0.019	1.62 (0.99–2.66)	0.055
SABA canisters
0–<3 (reference)	1.00		1.00
3–<12	1.44 (1.08–1.93)	0.014	1.37 (0.99–1.89)	0.055
≥12	1.65 (0.99–2.75)	0.057	1.78 (1.01–3.14)	0.046
Asthma-related ED visits
0 (reference)	1.00		1.00
≥1	3.76 (2.14–6.61)	0.000	2.62 (1.42–4.84)	0.002
Asthma-related hospitalisation
0 (reference)	1.00		1.00
≥1	3.19 (2.16–4.71)	0.000	2.59 (1.71–3.90)	0.000

Estimates tested by multivariable logistic regression analyses and reported as OR (95% CI). ICS: inhaled corticosteroid; LABA: long-acting β2-agonist; LAMA: long-acting muscarinic antagonist; LTRA: leukotriene receptor antagonist; SABA: short-acting β2-agonist; ED: emergency department. #: adjusted for all other factors in the model.

Specialist assessment waiting time

Among those without previous specialist contacts, 19% (1696 out of 8779) received specialist assessment within a 5-year follow-up period with a median (IQR) waiting time of 9 (2–28) months. As depicted in figure 3, we observed an increase in incident specialist assessments in the months shortly after inclusion as repeated OCS user. However, this effect declined after 6–8 months to a level appearing to be a baseline frequency of incident specialist referrals in the cohort.

FIGURE 3

Waiting time distribution for incident specialist assessment within 5 years of follow-up. OCS: oral corticosteroid; IQR: interquartile range.

Prescriber information

The majority of OCS dispensed during the baseline period was prescribed by general practitioners (71%) with an overall decrease from 79% in 1999 to 66% in 2018 (figure 4). Prescriptions by hospital physicians increased from 17% in 1999 to 26% in 2018 (a total relative increase of 65%). Hospital physicians were more likely to prescribe the second OCS prescription compared to the first prescription (analysis restricted to the years 2014–2018; figure 5). The amount of OCS prescriptions without prescriber source information reduced from 44% in 1999 to 6.0% in 2018 (not shown).

FIGURE 4

Prescriber source information on oral corticosteroid (OCS) prescriptions dispensed during the baseline period.

FIGURE 5

Prescriber source information distributed by a) first and b) second oral corticosteroid prescription, restricted to 2014–2018.

Sensitivity analyses and post hoc analyses

In the sensitivity analyses of patients with three and four annual OCS prescriptions, we found a slight increase in the proportion of previous specialist contacts from 22% (two OCS prescriptions) to 25% and 26% for three and four OCS prescriptions, respectively. However, the frequency of incident specialist contacts within 1 year of follow-up decreased slightly from 11% (two OCS prescriptions) to 9.9% (three OCS prescriptions) and 8.8% (four prescriptions).

The post hoc analysis restricted to patients with possible severe asthma showed similar trends of increasing referral for specialist assessment, but with larger fluctuations, which was probably due to the lower population number (supplementary figure S2).

Discussion

In this observational nationwide cohort study of young adults with asthma in Denmark, we found that patients with repeated OCS treatments are mainly managed in primary care. Overall, 70% of the patients did not have contacts to specialised care within either five prior to or 1 year post-inclusion. However, among those without previous specialist contacts, the frequency of incident specialist referrals tripled over the 20-year observation period from 6% to 18% a year. These results illustrate an opportunity for a potential optimisation of the referral pathway for patients with uncontrolled asthma who are at risk of long-term treatment side-effects [9, 17, 19]. While previous studies have mainly focused on specialist referrals among severe asthma populations [3, 4, 14, 15], the aim of this study was to explore trends and tendencies in a general asthma population with a specific focus on repeated OCS treatments as a “red flag” for identification of at-risk patients. Patients with repeated OCS use are at risk of both long-term morbidity as well as underestimation of the true severity of the disease. As stated in a recent national report from the United Kingdom, many cases of death due to asthma occurs in seemingly mild–moderate cases, highlighting a potential undertreatment of the disease [28]. Implementation of repeated OCS use as an easy-to-recall criterion and indication of specialist referral would be relevant for both primary care and hospital physicians.

Our study parallels findings from previous studies on uncontrolled and severe asthma populations which have also found a potential room for improvement in the overall patient referral pathway. A Danish cross-sectional study from 2014 found that only 14% of patients with low asthma control had contact to a respiratory specialist within 365 days [3]. Among patients with severe asthma and low control, the number was somewhat higher at 36% [3]. A more recent Danish study from 2021 found that 61% of patients with possible severe asthma were exclusively managed in primary care during 2014–2018 with significant differences in socioeconomic parameters compared to those achieving specialist referral [14]. Similar trends have been found in other countries. In Sweden, a register-based study found that only 20% of severe asthma patients were managed in secondary care [4]. Furthermore, only 32% of severe asthma patients had an asthma-related primary care contact within 1 year of inclusion, which indicates an overall low frequency of asthma-related healthcare contacts among severe asthma patients [4]. Studies from the United Kingdom have found similar trends, with only a minority of patients with uncontrolled and severe asthma being referred for specialist care [15, 29, 30]. Bloom et al. [29] found that the prevalence of asthma patients receiving three or more OCS courses a year had increased from 1% in 2006 to 2% in 2016, and that generally <20% of the patients were referred for specialist care. Encouragingly, one interesting finding of the study was that the specialist referral rates of eligible patients continually increased, which is consistent with our findings.

We furthermore found that acute asthma-related hospital visits and dispensing ≥12 SABA canisters were independent predictors of receiving specialist assessment, in agreement with previous literature [29]. This indicates that patients with difficult-to-treat and possible severe asthma are being referred to specialist assessments to a greater extent in agreement with current recommendations [8]. In the crude analysis, older age and residency outside the Capital region were associated with lower odds of specialist care, as found by a previous Danish study [14].

In average, 71% of the OCS was prescribed by general practitioners, which is lower than the 76% found in an Australian asthma study [31], but higher than the 60% found in a recent German study [32]. Interestingly, this proportion decreased over the study period as a further indication that more patients requiring OCS treatments are being managed in specialised care. This may be due to changes in asthma guidelines or increased implementation thereof; however, exploration of such underlying reasons was beyond the capability of this study.

Clinical considerations

Our results indicate that many patients with potentially uncontrolled asthma are not referred for specialist assessment. Repeated use of OCS is not considered an independent criterion for referral in, e.g. Danish, guidelines. However, it is recognised by international experts [11, 12] and stated in the GINA guidelines [8], as repeated OCS use indicates uncontrolled disease [11], and even a few courses is associated with significant adverse effects [9].

Uncontrolled asthma might be caused by difficult-to-treat asthma, i.e. lack of adherence, or severe asthma in need of medicine administered only by hospital specialists. In both cases, to prevent long-term complications of uncontrolled asthma, timely referral of at-risk patients is essential [12, 33] and easy-to-recall indicators are warranted. Specialist care for at-risk patients is associated with improved asthma-related outcomes [34, 35]. A national report from the United Kingdom found that 19% of asthma deaths were potentially attributable to a lack of specialist referrals [28]. Specialists might identify treatable traits such type 2 inflammation, and address comorbidities such as bronchiectasis, inducible laryngeal obstruction, heart diseases, allergic bronchopulmonary aspergillosis and eosinophilic granulomatous polyangiitis. Additionally, biological treatment for severe asthma, which has been proven to reduce both exacerbation rates and maintenance OCS use [36], is only available through hospital care in Denmark. While only 800–900 patients receive biological treatment in Denmark, it is estimated by experts that 10 000 may be eligible for this treatment [37], underpinning that more patients with OCS use could benefit from referral. One way to ease this process could be implementation of an easy-to-recall recommendation of referral for patients in need of more than one OCS treatment within a year. Digital applications and computerised decision support systems may further be of aid, as well as formal collaborations with pharmacists.

Strengths and limitations

The nationwide Danish registers provide data on all individuals residing in Denmark and are generally of high validity and completeness with the opportunity of data-linkage on an individual level [38]. They provide real-world data, which are collected systematically and independently of the researchers.

There are several important limitations to this study. First, due to the lack of diagnostic data from general practice and the low positive predictive value of asthma diagnoses in the National Patient Register [39], we constructed an asthma cohort based on validated methods using prescription data [25, 26]. The approach required a strict upper age of 45 years to limit the inclusion of patients with COPD. This limits the generalisability of the results to older asthma populations. Second, relevant clinical information on measures such as smoking, body mass index, spirometry parameters and indications for prescribed treatment were not available. We sought to limit including OCS use due to other reasons than asthma by censoring patients with potential OCS-treated comorbidities. However, we cannot account for other potential clinical factors contributing to the OCS use, such as allergies. Third, a dispensed prescription is not necessarily equal to the medication amount taken, and we are not able to account for possible stockpiling. However, the use of dispensed prescriptions did reduce the risk of misclassification due to primary nonadherence. Fourth, we did not have information on asthma severity, as dispensed asthma medication is no longer recommended for imputing asthma severity in observational studies [8], nor did we have data on asthma endotypes, hence limiting the identification of possible candidates for biological treatment. Finally, our definition of “specialist assessment” was not restricted to physicians with a speciality in respiratory medicine, as we considered this definition too restrictive for the purpose of this study.

Despite the noted limitations, we expect that the used definitions and chosen analyses have revealed results which reflect to a high degree the actual trends and predictors of specialist assessments among patients with repeated OCS use.

Conclusion

The proportion of patients being referred for specialist assessment has increased markedly over the past two decades; however, only 30% of adults with asthma and repeated OCS use are managed in specialist care overall. Although clarification of underlying reasons and/or barriers for most patients not achieving specialist assessment was beyond the capability of this study, our findings call for focusing on and optimisation of the patient referral pathway for high-risk patients with poor asthma control. Repeated use of OCS may serve as an easy-to-recall red flag for identification of patients with uncontrolled asthma where specialist referral should be considered. Future studies should focus on the feasibility of implementing this recommendation as an intervention in randomised controlled studies to assess whether patients referred to specialists on behalf of a red flag signal may benefit in form of faster assessment and better overall asthma management. In addition, studies should focus on identifying potential barriers of referral and exploring other instruments for optimising the complex patient pathway.

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material 00142-2022.SUPPLEMENT