The Fear of Asthma Symptoms Scale and the Asthma Behavior Checklist: preliminary validity of two novel patient reported outcome measures

Abstract

Objective: The aim of this study was to investigate the preliminary validity of two novel scales, the Fear of Asthma Symptoms scale (FAS) and the Asthma Behavior Checklist (ABC). Methods: Using cross-sectional design, data was collected online from 188 adult participants (Age 18-71 years) with a diagnosis of asthma and self-reported anxiety related to asthma, recruited through social media. Confirmatory factor analysis, internal consistency and test-retest reliability were ascertained to address validity.Results: The confirmatory factor analysis demonstrated convergent validity for both the FAS (average variance extracted; AVE=.57) and the item-reduced ABC-8 (AVE=.61) as well as divergent validity for both scales. Both scales demonstrated high internal consistency (FAS: α = 0.94; ABC-8: α = 0.92). Test-retest reliability assessed after 1 week was good (FAS: r=.85; ABC-8: r=.88).Conclusions: We observed promising psychometric properties of the FAS and the ABC-8. The two novel scales could be useful to identify excessive fear and avoidance in patients with asthma and to investigate putative mechanisms in clinical research on anxiety related to asthma. Further evaluation of psychometric properties in independent samples are needed.

Keywords:

• Asthma anxiety
• avoidance behavior
• convergent validity
• divergent validity
• reliability
• psychometric properties

Introduction

Anxiety in asthma

The connection between asthma and anxiety is well-established in large population-studies (1,2). Anxiety increases the risk for poorly controlled asthma (3,4) and the risk for low quality of life (5). To inform treatment and assessment of these patients, detailed knowledge is needed about the anxiety-asthma connection (6). Anxiety is characterized by a behavioral pattern of fear and avoidance which has been identified as a successful treatment target in cognitive behavior therapy (CBT). It has recently been shown that CBT targeting fear and avoidance is feasible to improve anxiety related to asthma (7,8). Hence, a detailed assessment of fear and avoidance may be important to facilitate the identification of individuals who could benefit from CBT.

Fear and avoidance in anxiety related to asthma

Even though attention to symptoms may be adaptive in asthma, anxiety could lead to hypervigilance (i.e. excessive attention to bodily symptoms similar to asthma) (9). Hypervigilance may cause individuals with anxiety and asthma to misinterpret safe bodily sensations as dangerous signs of asthma (10). The individual with asthma and anxiety may thus report more symptoms and experience an increasing number of situations as dangerous, which in turn may lead to a pattern of excessive avoidance (11). Excessive fear and avoidance of asthma symptoms may partly explain the low quality of life previously reported in individuals with asthma and anxiety (5). There is currently a scarcity of validated scales specifically assessing fear of asthma symptoms and avoidance behavior due to anxiety related to asthma.

Assessing fear and avoidance in asthma

Anxiety sensitivity has been shown to be relevant to asthma-related anxiety, and the Anxiety Sensitivity Index (ASI) has been used in studies on anxiety and asthma to assess that connection (12). However, the ASI is a broad measure specifically developed to assess fear of symptoms of anixety (13). Thus, even though the ASI may be able to identify persons at risk for anxiety disorders in the asthma population, it may not be able to assess the anxiety specifically related to asthma. Catastrophizing thoughts, such as “I don´t get enough air” are common during asthma exacerbations, whereas anxiety related to asthma may also be driven by an ever-present fear of asthma symptoms that triggers avoidance (14). The Catastrophizing about Asthma Scale (CAS) (15) assesses intense cognitions related to asthma exacerbations (i.e. the sensation of dsypnea) (15). However, the CAS may not fully capture fear related to hypervigilance, a factor that has been shown to be important in other somatic disorders affected by anxiety (16). The CAS has previously only been validated in one small studie (N = 92) including adults with mild to moderate asthma. Hence, CAS is not a gold standard instrument but, to the best of our knowledge, the only instrument available in the area. There is thus a need to develop and validate an instrument that could assess a broader and more persistent fear about asthma. There is a growing scientific interest to further investigate the role that avoidance behaviors play in anxiety (17), which highlights the need for valid instruments assessing these factors. To the best of our knowledge, there is yet no validated measurement that assesses excessive avoidance behavior in asthma. A previous study demonstrated that an online CBT program targeting excessive avoidance may reduce catastrophizing about asthma as well as improve perceived asthma control (8). For this study two scales were developed and tested to assess fear of asthma symptoms and avoidance behavior due to asthma. However, to properly investigate the pathways in the treatment of anxiety related to asthma, there is a need to validate the scales. Furthermore, fear and avoidance in asthma may be important constructs to assess for identification of patients that could benefit from psychological treatment (18,19).

For a scale to be validated there is a need for a series of analyses on data collected from the target population (20,21). A factor analysis investigates how large the common variance is between items and how many subscales the scale consists of. Tests of validity need to be performed, where convergent validity describes how well the items seem to assess the latent variable and divergent validity how the scale is different from other scales. Internal consistency reliability is a measure of the interrelationship of included items and to which degree the items jointly assesses the construct. Test-retest reliability examines the stability of a scale when administered at different time points. The aim of this study was to conduct preliminary validation of two novel scales: the Fear of Asthma Symptoms Scale (FAS) and the Asthma Behavior Checklist (ABC).

Methods

Participants

This study used a cross-sectional design collecting data from 188 participants (age 18–71 years) with asthma and anxiety related to asthma through advertisements on social media between June 2019 to September 2020.

Procedure

Participants were recruited via Facebook ads and followed a link to a secure database where information about the study and digital informed consent was provided. All measures were completed online, without any contact with the study staff. The survey contained demographic data, self-report on somatic and psychiatric diagnoses and the scales used for validation. All included participants confirmed having received an asthma diagnosis by a physician. Participants also reported if they had some anxiety about their asthma (“Do you feel stressed or worried about your asthma?” yes/no) and/or if they felt restricted by asthma, which could indicate a pattern of avoidance behavior (“Do you feel hindered by your asthma in your daily life?” yes/no). Estimated time required for the assessment was 15 min. To assess test-retest reliability, participants were invited to answer a repeated assessment after one week. Individuals who did not answer the first round of questions were excluded. All participants were offered a 100 SEK (≈ $11.50) voucher. We used a secure database (BASS4) that is developed to collect and store digital data in clinical research at Karolinska Institutet, Stockholm, Sweden. Ethical approval was received from the Swedish Ethical Review Authority in May 2019 (ID: 2019–02947).

Measures

Item generation and item reduction in the validated scales

The Fear of asthma symptom scale (FAS) was based upon a scale measuring hypersensitivity to visceral symptoms in irritable bowel syndrome, the Visceral Sensitivity Index (VSI) (22). The item selection of the FAS was discussed within the research group which consisted of experts within the field of asthma and CBT. Included items describe ongoing fear of asthma symptoms, such as “I am constantly aware of changes in my breathing” and “I worry that when I leave home, I will experience asthma symptoms”. The feasibility of the preliminary version of FAS was tested in a study evaluating online-CBT for anxiety related to asthma (8). The internal consistency in the sample was high (Cronbach’s alpha =0.83). After this preliminary evaluation, 1 item was omitted as it was deemed to assess a different construct (avoidance behavior). The version of FAS validated in the current study thus comprise 12 items, with response options from 0 (Do not agree at all) to 4 (Very much agree), see Table 1 for the full scale.

Table 1. The fear of asthma symptoms scale (FAS) with the initial 12 items.

FAS Below you find statements that describe how some people may feel about symptoms related to asthma. Symptoms of asthma may include shortness of breath, cough, wheezing or pressure in the chest. Specify how much you agree or disagree with each statement (0 = Do not agree at all, 5 = Very much agree). I worry that I will feel symptoms of asthma when I leave home. I get worried if I visit a place I have not been to before. I often worry about asthma. I have a hard time having fun because I can’t stop thinking about my asthma symptoms. I am often afraid that I will not be able to have a normal lung function. As soon as I sense symptoms of asthma, I get worried and anxious. I am constantly aware of my breathing. I often think about that asthma is a serious disease. As soon as I wake up, I start to worry that I will have trouble with my asthma during the day. When I feel stressed, I get worried about asthma. I keep thinking about what’s going on with my breathing. When I sense symptoms of asthma, I get anxious.

The Asthma Behavior Checklist (ABC) was first developed for clinical use in a CBT-trial for anxiety related to asthma (7). Items were generated from the excess behaviors (i.e. behaviors used too often or in the wrong occasions) that participants described in the initial trial, described in detail elsewhere (7). The technical construction of the scale was based upon the IBS-BRQ that measure avoidance in irritable bowel syndrome (23). Example of included items are “(Due to my asthma) I often check my breathing”, “(Due to my asthma) I avoid taking walks with other people”. The list of items was reviewed by psychologists with extensive experience in CBT for somatic conditions. This resulted in a scale of 31 items tested in the online-CBT trial for anxiety related to asthma, demonstrating high internal consistency in the sample (Cronbach’s alpha =0.90) (8). After this study two items were removed from the ABC. One item “(Due to my asthma) I avoid fur animals/smoke/perfume, even though I am not allergic” may have targeted an adaptive avoidance, and most participants in our studies have had allergies. The other item “(Due to my asthma) I avoid thinking about my breathing” had a very low correlation with the full scale (r= −0.11), maybe due to the vague cognitive avoidance which did not match well with other overt avoidance behaviors in the scale. The version of ABC administered in this study thus comprised 29 items. The response options were on a 7-graded scale, from 1 (never) to 7 (always), see Table 2 for the full scale.

Table 2. The Avoidance Behavior Checklist (ABC) with the initial 29 items.

ABC Here is a list of common behaviors that many people with asthma do to reduce or avoid asthma symptoms. Indicate how well each statement below is true for you, from 1(Never) to 7 (Always). I prefer not taking the stairs I avoid hurrying I avoid exercise or physical strain I exercise, but less vigorous than I wish to I prefer not to take the bus or the subway when I feel symptoms of asthma. Whenever I can, I avoid crowds I take the stairs, but avoid walking fast or running I try to do calm breathing I often check my breathing I try not to cough out of fear of not being able to stop When I'm with people I do not know very well, I try not to laugh out of fear that it will turn into a cough I avoid taking walks I go for walks, but preferably not with others I avoid going to the cinema / theater / concerts / lectures I go to the cinema / theater / concerts / lectures but prefer to place myself so that I can easily leave the room. When I feel symptoms of asthma, I prefer not to talk in meetings or in other similar contexts. I deliberately breathe through my nose to control my asthma I make sure to breathe through a scarf or other face protection when the weather is cold. I think about taking it easy when I am with children, grandchildren or pets. I do not play wild or strenuous games with children, grandchildren or pets. I avoid becoming short of breath I try to do things as calmly as I can I bring a water bottle or throat tablets I plan my route so I will not have to walk I rest before activities I avoid having sex I hold back, not to exert myself too much I prefer not to travel long distances by train, airplane, or bus I avoid being far from available healthcare

Measures to assess divergent validity

Catastrophizing about Asthma (CAS) comprise 24 items assessing catastrophizing thoughts during an asthma attack and in daily life rated from 0 (not at all) to 4 (certainly) (15). Examples of items are “During an asthma attack I worry all the time whether the asthma attack will end” and “In general, when I do not have an attack, I feel I can´t stand it anymore”. The CAS is derived from the Pain catastrophizing scale (PCA) (24), and has been validated in a clinical sample (N = 92) demonstrating good psychometric properties with a high internal consistency (Cronbach’s alpha = 0.93) and excellent test-retest reliability (r = 0.94), as well as a moderate correlation (r=.30) with perceptions of asthma symptoms (15). The CAS was independently translated and back-translated by two of the authors prior to the study, after which the translation was double-checked by a native English-speaking professional editor. Asthma control test (ACT) (25) is an often used questionnaire to assess perceived asthma control, consisting of five items. ACT has shown good internal consistency (Cronbach’s α =0.84) and is acceptable to specify degree of asthma severity (26). Anxiety Sensitivity Index (ASI) is a validated 16-item scale to measure fear of anxiety symptoms (13). The response options range from very little (0 points) to very much (4 points) and include items such as “It scares me when I feel shaky” and “It is important to me to stay in control of my emotions” (13)². Established translations to Swedish of both the ACT and the ASI were used (27,28).

Statistics

Data analysis was conducted in R, version 4.1.0 (29). using the lavaan package (30) in a confirmatory factor analysis (CFA). To test the validity of the FAS and ABC scales, a multi-factor CFA was conducted including the FAS, ABC, ACT, ASI, and CAS scales. Items with an R2 below 0.40 were removed after qualitative inspection (e.g. the item may be dubious to the responder or describe an unusual event) (21). The multi-factor CFA allowed for studying both convergent and divergent validity within the same model. For convergent validity averaged variance extracted (AVE) > 0.50 was used (31). AVE is a measure to investigate how much of the variance can be explained by the latent variable in relation to the amount of error variance. Divergent validity was demonstrated if the correlation between scales was smaller than the correlation within scales (i.e. the square root of the AVE). Divergent validity was also checked by calculating a Pearson correlation coefficient on the scale variables. Internal consistency was analyzed with Cronbach’s alpha. All participants who completed the second administration of questionnaires (N = 172), were included in a test-retest analysis using Pearson correlation.

Results

Clinical characteristics

The sample was characterized by a large proportion of females (n = 156; 83%), a mean age of 39.4 years (SD = 15.75) and, having had a diagnosis of asthma for a mean duration of 18 years (SD = 14.12), see Table 3 for full details. Several participants had received a psychiatric diagnosis at some point in their lives, of which depression (n = 74, 39%) and anxiety disorders (n = 55; 29%) were the most common. A majority reported worry about their asthma n = 128; 68%) or that daily life was restricted due to asthma (n = 122, 65%). Many individuals indicated that they had a chronic somatic syndrome, and numerous autoimmune diagnoses were present in the sample.

Table 3. Sample characteristics (N = 188).

	N (%)
Female	156 (83%)
Age, mean (SD)	39.4 (15.75)
Education
High school	82 (44%)
University	79 (42%)
Other^a	27 (14%)
Employment^b
Full time	69 (37%)
Part time	44 (23%)
Student	47 (25%)
Retired	26 (14%)
Unemployed	16 (9%)
Asthma^c
Duration years, Mean (SD)	18.61 (14.12)
Controlled asthma (ACT ≥ 20)	67 (36%)
Partially controlled asthma (ACT 16-19)	61 (33%)
Uncontrolled asthma (ACT ≤15)	58 (31%)
Psychiatric diagnoses
Depression	74 (39%)
Anxiety disorder	55 (29%)
Other^d	45 (24%)
Chronic somatic syndromes
Allergies	150 (92%)
Eczema	53 (28%)
Auto-immune disease^e	23 (12%)
Pain disorders^f	35 (18%)

a Elementary School, Vocational School, or not specified. ^bThe total sum exceeds 188, since some participants worked parttime and studied or were part-time retired. ^cAll participants answered demographics about asthma, but 2 were missing from the Asthma Control Scale. ^dADHD, Autism, OCD, Eating disorder, Personality disorder. ^eUlcer Colitis, Chron’s Disease, Diabetes, Hypothyroid, Celiac Disease, Alopecia Universalis, Psoriasis, Rheumatism, Sjogrens Syndrome. ^fFibromyalgia, Chronic Pain, Irritable Bowel Syndrome.*

Confirmatory factor analysis

A multi-factor CFA model with all 5 scales was constructed to test convergent and divergent validity. In the FAS, item 2 had an R2 below 0.4 and 11 items were thus retained. However, for the ABC many items were removed from the scale (items 4– 20, 23, 26, 28 and 29), leaving 8 items loading sufficiently on the latent variable. We will refer to this shortened scale as ABC-8. Qualitative inspection revealed that items with low factor loadings may have been ambiguous, describe a behavior that is unusual or describe a common habit which most patients with asthma use. See Tables 4 for the validated scales with factor loadings.

Table 4. The validated versions of the Fear of Asthma Symptom scale (FAS) and the Asthma Behavior Checklist (ABC).

FAS	R^2	Item description	Mean	SD
Item 1	0.61	Worry about asthma when leaving home	1.67	1.7
Item 2	0.69	Often worry about asthma	2.34	1.6
Item 3	0.48	Difficult having a good time due to worry about asthma	0.92	1.3
Item 4	0.59	Worry about lung function	2.52	1.8
Item 5	0.61	Sense asthma symptoms, become anxious	2.94	1.7
Item 6	0.54	Constantly aware of breath	2.41	1.6
Item 7	0.54	Often think asthma is a serious disease	2.11	1.7
Item 8	0.55	Anxiety for asthma during the day when waking up	0.90	1.4
Item 9	0.62	When in stress, anxious about asthma	2.06	1.8
Item 10	0.57	Constantly aware of breath changes	1.71	1.6
Item 11	0.52	Asthma symptoms cause fear	2.61	1.7
ABC-8	R^2	Item description	Mean	SD
Item 1	0.48	Prefer not taking the stairs	3.0	1.9
Item 2	0.46	Avoid hurrying	3.3	1.8
Item 3	0.56	Avoid exercise	3.1	1.9
Item 4	0.71	Avoid becoming short of breath	3.5	2.1
Item 5	0.76	Try to do things as calmly as I can	3.2	2.0
Item 6	0.59	Plan my route to avoid walking	2.3	1.8
Item 7	0.58	Rest before activities	3.1	2.1
Item 8	0.75	Hold back not to exert myself	3.2	1.9

Note. FAS: Response range 0–4 (“Do not agree at all” – “Very much agree”).

ABC: Response range 1–7 (“Never” – “Always”).

R² = squared standardized loading on the latent variable.

Convergent validity

Both the FAS (AVE=.57) and the ABC-8 (AVE=.61) had an acceptable convergence (AVE > 0.50) as measured in the full model, see Table 5 for AVE on all scales in the model.

Table 5. Convergent validity analyzed with average variance extracted (AVE).

Measure	AVE
FAS	.57
ABC-8	.61
CAS	.51
ACT	.39
ASI	.41

Note. AVE > 0.50 is required for convergent validity.

FAS = Fear of Asthma Symptoms Scale. ABC-8 = Asthma.

Behavior Checklist. CAS = Catastrophizing about Asthma Scale.

ACT = Asthma Control Test. ASI = Anxiety Sensitivity Index.

Divergent validity

The FAS demonstrated divergent validity with all other scales, but CAS (r=.87). The ABC-8 demonstrated divergent validity, the correlations between all other scales (r = 0.58, -. 65) was smaller than the item-correlation within the scale (r=.78), see Table 6.

Table 6. Correlations between scales versus the squareroot of AVE.

	FAS	ABC-8	√AVE
FAS		.65	.76
ABC-8	.65		.78
CAS	.87	.60	.71
ACT	-.71	-.61	.63
ASI	.61	.58	.64

Note. AVE = average variance extracted. FAS = Fear of asthma symptoms scale.

ABC-8 = Asthma behavior checklist. CAS = Catastrophizing about asthma scale.

ACT = Asthma control test. ASI = Anxiety sensitivity index.

Internal consistency and test-retest reliability

Both the FAS and the ABC-8 showed high internal consistency (FAS α = 0.94 [CI 95%: 0.92, 0.95]; ABC-8 α = 0.92 [CI 95%: 0.90, 0.94]), and good test-retest reliability (correlations of scale scores over time of r = 0.85 and r = 0.88 respectively).

Bivariate correlations

Bivariate correlations between all scales demonstrated that the FAS was strongly correlated with the CAS (r=.83), and ABC showed a strong correlation with the FAS (r = 0.61), see Table 7 for full details. The correlation with ASI was moderate for both the FAS (r=.57) and the ABC (r=.54). The ACT correlated moderately with both the FAS (r=.55), the ABC (r=.42) and the CAS (0.55) and weakly with the ASI (r=.24). Both the ABC and the FAS strongly correlated with the repeated assessment of each scale (T2). Furthermore, the FAS and the ABC strongly correlated both at time 1 (r=.61) and at time 2 (r=.68).

Table 7. Bivariate correlations.

	ABC -8 (N = 188)	FAS (N = 185)	ACT (N = 185)	ASI (N = 185)	CAS (N = 184)	ABC-8 t2 (N = 177)	FAS t2 (N = 173)
ABC −8		.61	-.42	.54	.56	.88	0.58
FAS	.61		-.55	.57	.83	.61	.85
ACT *	-.42	-.55		-.24	-.55	-.43	-.52
ASI	.54	.57	-.24		.61	.57	.59
CAS	.56	.83	-.55	.61		.59	.77

Note. FAS = Fear of Asthma Symptoms Scale. ABC-8 = Asthma Behavior Checklist. CAS = Catastrophizing About Asthma Scale. ACT = Asthma Control Test. ASI = Anxiety Sensitivity Index.

Bold figures show strong correlations (r= 0.60–0.79), bold and underlined figures very strong correlations (r= 0.80–1.0). * Higher rating on ACT means improved asthma control, while higher rating on all other scales signifies increased problem burden. The ACT is therefore expected to correlate negatively with all other scales. T2 = Repeated assessments after 1 week only on the FAS and the ABC for test-retest analyses.

Discussion

This was a cross-sectional study to preliminary validate two novel scales, the FAS and the ABC, on adults with anxiety related to asthma. All but one item in the FAS covaried well with the latent variable, while many items in the ABC were removed due to low factor loadings on the latent variable, leaving 8 items (ABC-8). Both the FAS and the ABC-8 demonstrated acceptable convergent and divergent validity. The internal consistency was high for both scales, and test-retest reliability good. Bivariate correlations demonstrated construct validity for the ABC-8, and for the FAS. Qualitative inspection of items with low factor loading in the ABC revealed that these may have been perceived as ambiguous (e.g. I take the stairs, but avoid walking fast or run), or describe a behavior that is unusual (e.g. I deliberately breathe through my nose to control my asthma). Other items with low factor loadings may describe a rule-governed habit, such as “I breathe through a scarf or other face protection when the weather is cold”, which may be a reasonable avoidance in the daily management of asthma. The items retained in the scale ABC-8 all describe more general avoidance behaviors. Notably, the FAS seems to share an overlapping construct with the CAS as they are highly correlated. In this sample on an asthma-population with anxiety related to asthma the FAS seem to fair better, with a higher average variance explained than the CAS. Future studies are needed to better understand if there are any true differences between constructs assessed with these scales, or if they are interchangeable. In the latter case, the FAS, being a 11-item questionnaire, might be a better choice as the CAS comprise 24 items. The ASI is developed and validated to investigate fear of anxiety symptoms (13). In this sample, the ASI did not show optimal convergent validity, even though it correlated quite strongly with the CAS (r = 0.61). As such, the ASI may not be ideal as an outcome to assess anxiety related to asthma. The Asthma Control Test (ACT) assesses self-perceived asthma control, our observation was that ACT did not correlate well with the ASI, but moderately well with the FAS, the ABC and the CAS. This might strengthen a previous suggested hypothesis that anxiety in asthma and other respiratory diseases is symptom-specific rather than related to a general fear of anxiety symptoms or a specific clinical anxiety disorder (32,33).

Limitations

The results in this study should be regarded as preliminary due to the limited sample size and recruitment of a convenience sample. Most participants were female, which is in line with previous research on internet-delivered clinical treatments for anxiety and depression (34). The lifetime prevalence of anxiety in this sample may seem high (29%) but is in line with prevalence previously reported in the general population (34%) (35). The results may therefore mainly be generalized to females with anxiety related to asthma. Furthermore, we had to shorten the ABC quite substantially. Consequently, further studies are needed to validate the ABC. Importantly, even though the sample size is limited it is probably not underpowered. Mundfrom et al. investigated the sample size needed for conducting factor analyses based on a variables (items)/factor ratio in a simulation study and found that with a variables-to-factors ratio of at least 7 the minimum necessary sample size for excellent agreement is never greater than 180 (36). Because we observed a one-factor solution on both scales, and variables within those factors were 12 and 8 respectively, it seems that this study was correctly powered. Also, factors that lower the required sample size include high communalities (> = 0.40), high factor loadings (> = 0.50), and the number of high loading items (> = 5) per factor. Given that this was the case, we do not believe our study was lacking power (37). Nevertheless, further and larger studies are needed to ensure the psychometric qualities of the FAS and the ABC. Future studies should decide dimensional validity and norms for both scales including healthy controls and clinically recruited patients with asthma, also including patients without anxiety related to asthma. Furthermore, sensitivity to change due to intervention need to be estimated in a randomized controlled study that control for the effect of time.

Conclusions

Both novel scales, the FAS and the ABC-8, demonstrate promising psychometric properties and could be useful to assess anxiety related to asthma. The results need to be confirmed in larger studies that also include healthy controls to decide dimensional validity and clinical cut-off scores.

Declaration of interest

The authors declare that there are no conflicts of interest.

Additional information

Funding

The study was funded by the Swedish Research Council for Health, Working life, and Welfare (FORTE, Grant No 2015–00289), the Swedish Heart-Lung Foundation (Grant No 20180512 and 20210416), the Swedish Asthma and Allergy Association Research Fund (F2018-0034, K2020-0008 and F2021-0012) and the Swedish Research Council (Grant No 2018–02640). None of the funding bodies had any influence on study design, implementation, analysis, or interpretation. This work used services from the eHealth Core Facility at Karolinska Institutet, which is supported by the Strategic Research Area Healthcare Science (SFO-V).