Leventhal’s common-sense model and asthma control: Do illness representations predict success of an asthma rehabilitation?

Abstract

This study investigates whether illness representations predict changes in asthma control or vice versa. Illness perceptions and asthma control were assessed in N = 113 asthma patients at the begin/end of an inpatient rehabilitation. Bivariate cross-lagged panel analyses showed that the illness representation dimensions Personal control, Consequences, Coherence, and Emotional representation predicted the change in asthma control, but asthma control did not predict illness representations. In multiple regression analyses with covariates, Personal control still predicted the change in asthma control. Illness beliefs assessed at the beginning of an intervention might help to identify patients with lower probability of treatment success.

Keywords

asthma control change cross-lagged panel analysis illness representation prediction pulmonary rehabilitation

Introduction

Leventhal’s common-sense model (CSM) of illness representations (Leventhal et al., 2012) has received growing attention over the last 20 years in asthma research (Heyduck et al., 2015; Kaptein et al., 2010). This model describes how people react to health threats by forming both cognitive and emotional responses. Cognitive illness representations include perceptions about identity (for example labels and/or symptoms associated with the illness), timeline, cause, coherence consequences, and control of the illness, while emotional illness representations capture feelings as depression, anxiety, or anger associated with the illness (Diefenbach and Leventhal, 1996). The CSM postulates that cognitive and emotional representations are distinct but interrelated concepts and both may evoke health-related behavior such as coping or adherence. The results of these actions are evaluated again and may change cognitive and emotional illness representations and so on. Illness representations have been shown to associate with health-related behavior and health outcomes in a variety of chronic conditions, for example, fibromyalgia (Glattacker et al., 2010), type 2 diabetes (Searle et al., 2007), chronic back pain (Glattacker et al., 2013a), and mental illnesses (Baines and Wittkowski, 2013; Glattacker et al., 2013b).

Research on illness representations in asthma has shown relationships with a variety of outcomes, such as the number of hospitalizations, quality of life, and adherence (Byer and Myers, 2000; Horne and Weinman, 2002; Kaptein et al., 2008, 2010; Sofianou et al., 2013). For example, Jessop and Rutter (2003) found that beliefs concerning the identity, cause, and control of asthma predicted current adherence while future adherence was predicted by beliefs of control and perceived duration.

While associations between illness representations and adherence have been studied intensively, only few studies have investigated relationships between illness representations and asthma control. Asthma control is based on four criteria (daytime symptoms, symptoms at night, reliever need, and activity limitation due to asthma) and is regarded as the main treatment outcome in asthma (Boulet et al., 2015; Global Initiative for Asthma (GINA), 2014; Pedersen, 2010). However, asthma is often poorly controlled despite continuous improvement of asthma treatment (Demoly et al., 2009). Therefore, it is important to identify potential predictors of asthma control and to assess the causal pathways between these variables. Cross-sectional studies showed that illness representations may be candidates for causal factors. For example, Powell et al. (2011) showed significant zero-order correlations between asthma control and the illness representation dimensions Consequences, Personal control, Identity, Concern, and Emotional representation. After adjusting for other variables, Emotional representation still was associated with asthma control. In a study with adolescents, Tiggelman et al. (2014) identified Personal control as significant predictor in cross-sectional models.

However, cross-sectional studies must be complemented by longitudinal and experimental studies to investigate the causal role of illness representations. Negative or dysfunctional illness perceptions, for example, perceived low control, negative consequences, and acute timeline, may have a negative effect on asthma control over time. To our knowledge, only one longitudinal study investigated whether illness representations predict the change in asthma control but found no associations over a period of 2 years (Tiggelman et al., 2014). However, no study examined whether illness representations predict short-term changes in asthma control.

The CSM is theorized to be a dynamic model. Positive illness representations (for example, high controllability) at the beginning of an asthma rehabilitation may facilitate the change in asthma control (Burgess et al., 2015). However, the effects in the opposite direction are also conceivable, such that asthma control at the beginning of an intervention predicts the change in illness representations. For example, patients with low asthma control at the beginning of an intervention may have formed stable negative cognitions such as “My asthma is much worse than the asthma of other patients and is not controllable” and may therefore also show more unfavorable illness representations at the end of rehabilitation. No study has yet explored such bidirectional effects.

Therefore, this study investigates the following research questions: (1) Are illness representations in asthma patients at the begin and end of inpatient rehabilitation associated with asthma control? (2) Does asthma control at the begin of an inpatient rehabilitation predict short-term changes in illness representations? (3) Do illness representations at the begin of an inpatient rehabilitation predict short-term changes in asthma control? If this turned out to be the case, do cognitive and emotional representations predict short-term changes in asthma control independently?

Methods

Sample

This is a secondary analysis of a study that psychometrically tested the Health Education Impact Questionnaire (heiQ^™) in patients with different chronic conditions (Schuler et al., 2013, 2014). The project was approved by the ethical review committee of Hannover Medical School (No. 5070). All (N = 119) participants with asthma attending inpatient pulmonary rehabilitation in the rehabilitation center “Klinik Bad Reichenhall” between December 2009 and March 2010 were included; six patients did not complete inpatient rehabilitation and were excluded. Details of the intervention were described elsewhere (Lingner et al., 2015). Both at the beginning (T1) and the end (T2) of rehabilitation, patients filled out several questionnaires including the short version of the Illness Perception Questionnaire—Revised (IPQ-R) and the Asthma Control Test (ACT).

Instruments

IPQ-R—short version

The IPQ-R (Glattacker et al., 2009; Moss-Morris et al., 2002) captures cognitive and emotional illness representations with the following seven subscales: Identity, Timeline, Consequences, Personal control, Coherence, Cycle, and Emotional representation. In this study, the subscale Identity was not assessed. The short version uses three items to assess each of the other six scales. Scale values range from 1 to 5. Higher values indicate more chronic (Timeline), more Consequences, higher Personal control, better understanding (Coherence), more cyclic occurrence of symptoms (Cycle), and higher emotional distress. Cronbach’s alpha coefficients (in our study) range from 0.48 (Coherence, T1) to 0.91 (Emotional representation, T2).

ACT

The ACT (Schatz et al., 2006) assesses asthma control with the following five items: (1) activity limitation, (2) shortness of breath, (3) awaking due to asthma symptoms, (4) reliever medication, and (5) global judgment of asthma control (perceived asthma control). All items refer to the last 4 weeks. They were scaled from 1 to 5 and the sum score indicates asthma control with values of 25/20–24/<19 meaning perfectly controlled/well controlled/poorly controlled asthma, respectively (Jia et al., 2013). Reliability estimates range from 0.77 to 0.85 (Schatz et al., 2006).

Statistical analysis

The mean changes in illness representations and asthma control were tested via paired t-test. To estimate the amount of change, the standardized effect size (SES) was computed with difference of the mean values (M_T2 − M_T1) as the numerator and the standard deviation of T1 as the denominator (Kazis et al., 1989). Correlations between Illness representations (at T1 and T2, respectively) were estimated for latent variables.

To analyze all three research questions at once, six cross-lagged panel models (CLPMs) were computed to test the relationships between each subscale of the IPQ-R and asthma control (Selig and Little, 2012). All models were computed in Mplus V7.2 with weighted least square estimator (WLSMV estimator) (Muthén and Muthén, 2015). The constructs were modeled as the latent variables. All items of the respective scale served as manifest indicators. Figure 1 illustrates the CLPM between an illness representation construct and asthma control. Research question 1 was tested via the cross-lagged paths cl1 and cl2 in Figure 1. Factor loadings and thresholds of an item were held constant across both measurement occasions, thus assuming strong factorial invariance over time for each construct.

Figure 1.

Example of a cross-lagged panel model for asthma control (ACT) measured by five items and an illness perception dimension (IPQ) measured by three items. cl1: cross-lagged path between latent ACT T1 and latent IPQ dimension T2; cl2: cross-lagged path between latent IPQ dimension T1 and latent ACT T2; r1: correlation between latent ACT and latent IPQ dimension at T1; r_res2: correlation between latent ACT residual and latent IPQ dimension residual at T2; aup1: autoregressive path between latent ACT T1 and latent ACT T2; aup2: autoregressive path between latent IPQ dimension T1 and latent IPQ dimension T2.

Overall model fit was assessed using chi² test, comparative fit index (CFI), root mean square error of approximation (RMSEA), and weighted root mean square residual (WRMR) with CFI >0.95 (0.90), RMSEA <0.5 (0.8), and WRMR <1.0 indicating good (acceptable) model fit. Model modifications (e.g. relaxing fixed parameters) of poorly fitting models were based on substantive considerations as well as statistical parameters (expected parameter changes and modification indices).

To further explore whether cognitive and emotional illness representations have unique effects on the change in asthma control (research question 3), we computed a multiple regression model with asthma control T2 as the dependent variable and the following predictors: asthma control at T1, Emotional representation at T1, and the cognitive subscale with highest impact on asthma control T2 according to the results of the CLPMs. Age, sex, and FEV1 at T1 were also included as covariates. To further test whether emotional and cognitive representations have the same influence on the change in asthma control, a model with both regression parameters restricted to be equal was tested using the Wald test.

Multiple imputation was used to deal with the missing values. The numbers of cases with missing data ranged from 0 percent to 2 percent. Ten imputed data sets were created using the Markov chain Monte Carlo method (Van Buuren, 2012) implemented in SPSS V23 (IBM Corp., 2014). Pooled results across all imputations were reported with one exception: in Mplus, pooled p-values of the chi² test cannot be computed for WLSMV estimator (Asparouhov and Muthen, 2010). Therefore, p-values of chi² tests of one imputation are reported if p-values did not differ remarkably between the 10 imputed data sets. Alpha was set to 0.05 for all analyses.

Results

Sample

The sample consists of N = 113 patients (41.5% female; with a mean age 47.2 years (standard deviation (SD) = 10.5) years. The mean FEV1 percent predicted was 86.5 (SD = 23.0). The mean value of ACT at the begin of rehabilitation was 15.51 (SD = 5.02) and the mean FEV1 after bronchodilator was M = 3.03 (SD = 0.87). Further sample statistics are reported in Table 1 and Schuler et al. (2016).

Table 1.

Sample statistics.

	n (%)^a
Women	47 (41.5)
Asthma severity^b
Intermittent	2 (1.8)
Mild persistent	32 (28.3)
Moderate persistent	60 (53.1)
Severe persistent	17 (15.0)
Living with a partner	84 (74.3)
Occupation
Laborer	61 (54.0)
Clerk/civil servant	44 (39.8)
Other	7 (6.3)
Employment status
Working	97 (85.8)
Unemployed	10 (8.8)
Other (e.g. housewife)	6 (5.4)
	M (SD)
Age	47.2 (10.5)
FEV1 (L)
Before bronchodilator	2.80 (0.87)
After bronchodilator	3.03 (0.87)
FEV1 pred (%)
Before bronchodilator	86.5 (23.0)
After bronchodilator	93.2 (22.1)
VC (L)
Before bronchodilator	3.85 (0.97)
After bronchodilator	4.04 (0.94)
FEV1/VC (%)
Before bronchodilator	72.88 (13.28)
After bronchodilator	74.47 (11.96)
ACT
At the begin of inpatient rehabilitation	15.61 (5.02)

SD: standard deviation; ACT: Asthma Control Test; FEV1: forced expiratory volume in 1 s; VC: vital capacity; %pred = % of predicted.

Some categories did not sum up to n = 113 (100%) because of missing values.

According to GINA (2006).

Changes in illness perceptions and asthma control

In nearly all IPQ-R subscales, statistically significant changes were shown between T1 and T2 (except Cycle), albeit effect sizes were small (between SES = −0.03 (Cycle) and 0.32 (Emotional representation). However, considerable changes were observed in ACT (SES = 0.73). Detailed results are found in online supplement.

Relations between Illness representations

Correlations between illness representations at T1 range from r = 0.02 (Timeline and Emotional representation) to r = 0.69 (Personal control and Coherence). At T2, correlations range from r = 0.07 (Timeline and Coherence) to r = 0.83 (Consequences and Emotional representation). All correlations are found in the online supplement.

Relations between illness perceptions and asthma control

All models yield acceptable fit values (detailed results are shown in online supplement). In all models, two thresholds in item 5 of ACT (“overall asthma control”) had to be set free to achieve good model fit. Two models could only be identified by fixing one residual correlation of an IPQ-R-Item to 0. For the IPQ-R-Scale Coherence an additional set of five thresholds had to be set free to achieve good model fit. Exact values of all measurement models (factor loadings, thresholds) are available on request.

Estimates of the structural parameters of all models can be found in Table 2. Asthma control had no significant effect on the changes in any IPQ-R subscale. In contrast, the changes in asthma control are predicted by Consequences (β₂ = −0.35, p < 0.001), Personal control (β₂ = −0.41, p < 0.001), Coherence (β₂ = −0.23, p < 0.001), and Emotional representation (β₂ = −0.25, p < 0.001). These results confirm research question 3 stating that illness representations predict the change in asthma control, as opposed to research question 2 stating the reverse relation.

Table 2.

Standardized path coefficients in all six models—WLSMV—imputation.

	Cross-lagged paths				Correlations						Autoregressive paths
	cl1	p	cl2	p	r1	p	r _res2	p	r2	p	aup1	p	aup2	p
Timeline	−0.09	0.286	−0.05	0.513	0.13	0.189	−0.21	0.151	−0.13	0.223	0.52	<0.001	0.58	<0.001
Consequence	0.13	0.142	−0.35	<0.001	−0.44	<0.001	−0.28	0.008	−0.52	<0.001	0.37	<0.001	0.75	<0.001
Personal control	−0.08	0.444	0.41	<0.001	0.27	0.013	0.08	0.621	0.41	<0.001	0.42	<0.001	0.80	<0.001
Coherence	−0.06	0.458	0.23	<0.001	0.12	0.220	0.05	0.681	0.19	0.036	0.50	<0.001	0.68	<0.001
Cycle	−0.19	0.067	0.16	0.100	−0.29	0.003	−0.28	0.022	−0.44	<0.001	0.48	<0.001	0.58	<0.001
Emotional	−0.05	0.538	0.25	0.003	−0.25	0.006	−0.12	0.286	−0.33	<0.001	0.46	<0.001	0.65	<0.001

cl1: cross-lagged path between latent ACT T1 and latent IPQ dimension T2; cl2: cross-lagged path between latent IPQ dimension T1 and latent ACT T2; r1: correlation between latent ACT and latent IPQ dimension at T1; r2: correlation between latent ACT and latent IPQ dimension at T2; r_res2: correlation between latent ACT residual and latent IPQ dimension residual at T2; aup1: autoregressive path between latent ACT T1 and latent ACT T2; aup2: autoregressive path between latent IPQ dimension T1 and latent IPQ dimension T2.

Independent effects of emotional and cognitive illness representations on asthma control

First, a latent multiple regression model with ACT T2 as dependent variable and ACT T1 and Personal control, Coherence, and Consequences as predictors was computed. In this model, none of the three cognitive illness representation scales showed a statistically significant influence but Personal control showed by far the highest regression parameter and the lowest p-value (β = −0.415, p = 0.067). The correlations between these cognitive illness representations were between ( $| r | = | 0.33 | - | 0.71 |$ ), which indicates considerable overlap. Second, a latent multiple regression model with ACT T1, Emotional representation T1 and Personal control T1 as predictors of asthma control T2 showed at least acceptable fit values (chi² = 151.7 (113), CFI = 0.979, RMSEA = 0.055, WRMR = 0.751). Personal control T1 (β_pc = 0.39, p < 0.001) still predicted the change in asthma control while Emotional representation (β_er = −0.12, p < 0.212) did not. The correlation coefficient between both predictors is r = −0.40 (p < 0.001). Controlling for age, sex, and FEV1 T1 did not change this pattern (data not shown, available on request). In addition, the Wald test showed that a model with restricted regression parameter (β_pc = β_er) was statistical significantly worse than a model with freely estimated parameters (Wald = 27.37(1); p < 0.001).

Discussion

This is the first study that explored the longitudinal relationships between illness representations and asthma control using latent variables, that is, controlling for measurement error (Bollen, 1989). While asthma control and illness representations were correlated at both the begin and end of inpatient rehabilitation, we found that asthma control at T1 had no impact on the changes in illness representations. In contrast, illness representations of asthma patients at T1 predicted the change in asthma control at T2. Patients with a better understanding of their illness and who believe in their power to control their asthma showed greater improvements in asthma control than did other patients. On the contrary, patients who believe that their asthma has major consequences for their life and are more distressed showed smaller improvements in asthma control. Therefore, assessment of illness representations at the beginning of an intervention may be used to identify patients who may improve less in asthma control and modify their intervention accordingly (for example, by presenting more information or by using techniques to change dysfunctional cognitions).

To explore the relationships among different aspects of cognitive illness representations and asthma control, we computed a regression model on ACT T2 with those cognitive representations as predictors that showed significant cross-lagged relationships in the bivariate CLPMs (i.e. Personal control, Consequences, and Coherence). The results of this model and the medium to high correlation coefficients at T1 indicate considerable overlap between these three representations. People who do not think that they have control over their asthma (i.e. low Personal control) tend to have less understanding of it and to see more serious consequences for their life. Similar results were shown in other studies (Horne and Weinman, 2002; Jessop and Rutter, 2003; Tiggelman et al., 2014) although correlation coefficients in our study were higher, which may be explained by our use of latent variables. Positive changes in ACT seem to be best predicted by the aspect of Personal control. Although not reaching statistical significance in this model (which is explained by the overlapping content of all three predictors), the point estimate of the regression coefficient of Person control on ACT T2 remains remarkably constant over all tested models. However, this does not mean that other cognitive illness representation dimensions are negligible. It just highlights that Personal control best captures those aspects of cognitive representations that predict improvements in ACT T2 during inpatient rehabilitation.

Emotional representation predicted the changes in asthma control only in the bivariate model, but not when additionally controlling for cognitive representations, that is, Personal control. Personal control and Emotional representation show medium correlation coefficients in both our study and in Jessop and Rutter (2003). This result may be in line with cognitive theories of emotions, in which emotional states are (at least in part) the results of cognitive appraisal processes (Oatley and Johnson-Laird, 2014). Cognitions of low control over asthma symptoms may lead to more negative emotional asthma representations. Thus, Emotional representation may at least, in part, mediate the relationship between Personal control and asthma control and therefore shows no significant effect on asthma control in a multivariable regression model. Future studies should further investigate the relationship between cognitive and emotional illness representations.

Our results are in line with other studies that showed longitudinal relationships between illness representations and asthma outcomes. For example, Calfee et al. (2006) demonstrated that perceived asthma control has predictive value on emergency healthcare utilization over 1–4 years. Tiggelman et al. (2014) found in early adolescents significant relations between actual asthma control and the illness representations Personal control and Identity. But illness representations had no associations with the change in asthma control over a longer period of time (2 years) in their study. However, besides varying age groups and time frames, there is another important difference between this study and the work of Tiggelman et al. (2014): we examined whether illness representations at the begin of an intervention predicted the changes in asthma control that were probably (at least in part) caused by this intervention, while Tiggelman et al. examined relationships over time without a planned intervention between time points. Therefore, we showed that illness representations predict at least short-term gains in asthma control achieved during an intervention. Future studies should examine whether illness representations may also predict long-term intervention-gains in asthma control.

The levels of asthma control at T1 had no impact on the changes in illness representations in our study. Contrary to considerations, patients with high asthma control at T1 did not show greater improvement in ACT than did patients with lower asthma control at T1. However, a statistical issue should be taken into account: The small effect sizes and the high autoregressive path coefficients indicate that illness representations remain considerably stable in this study. In some scales, this may reflect ceiling effects indicated by high values at T1 (e.g. Timeline).

Although significant relations between illness perceptions and asthma control were found, the causal paths remain unclear. One causal path may include health-related behavior, for example, higher physical exercise during inpatient rehabilitation and/or better medication adherence. However, the results on the relationship between illness perceptions and use of medications are mixed: some studies showed substantive relationships (Jessop and Rutter, 2003; Sofianou et al., 2013), but some did not (Calfee et al., 2006). More adequate illness representations may also lead to better coping strategies such as information seeking or less worrying (Kaptein et al., 2010; Tiggelman et al., 2014). Another idea is that higher Personal control might reduce perceived stress, which, in turn, might improve asthma control (Lu et al., 2014). Future studies should focus on testing different causal path models that may explain these relationships.

The relationships between illness representations and the change in asthma control may depend on the intervention performed between T1 and T2. In this study, a complex intervention accomplished by a multi-professional team including for example sophisticated diagnostics, exercise, and patient education took place. Hence, pre–post differences in ACT were high. An interesting question for further studies would be whether interventions with other settings (outpatient vs. inpatient, length of time) or thematic priorities (e.g. exercise only) would result in similar patterns.

Limitations

This study has several limitations. First, there was no experimental manipulation of illness representations or asthma control. Therefore, causal claims about the associations between illness representations and asthma control cannot be made. Second, rigorous tests of the psychometric properties of the short version of the IPQ-R could not be performed. For example, to test whether a scale is unidimensional and invariant over time, a minimum of four items per scale are necessary, but the short version of the IPQ-R contains only three items per scale. Third, a clear separation of within and between effects is not possible in a CLPM with only two measurement occasions. For example, high autoregressive path coefficients between T1 and T2 only mean that the order of patients tends to remain stable over time. But whether most people show the changes in the same amount or very different changes is not clear. Other statistical models such as the random-intercept-CLPM (Hamaker et al., 2015) clearly separate within from between-person effects and thus overcome this limitation. However, at least three measurement occasions are necessary to identify these kinds of models. A fourth limitation of our study is that with only two measurement occasions, we could not test whether the changes between two time points (e.g. between T1 and T2) in one construct predict the changes between two later time points (e.g. between T2 and a follow-up). We have only tested whether the changes are predicted by a previous state measure. Fifth, the illness representation dimensions Identity and Causal beliefs were not assessed in this study. The dimension Identity showed significant correlations with adherence (Jessop and Rutter, 2003) and therefore may also have an influence on the change in asthma control. The sample size may be regarded as a further limitation of the study. With higher sample size, some additional paths may have been identified as statistically significant. And finally, to achieve acceptable model fit, two thresholds of ACT-item 5 were set free. This means that scalar measurement invariance (Meredith, 1993) over time was not achieved in ACT. Future studies should examine measurement invariance of ACT over time more extensively.

In summary, this study demonstrated that illness representations, especially Personal control, predict the changes in asthma control during an inpatient rehabilitation. By assessing existing illness representations at the begin of an intervention, healthcare providers might benefit in understanding possible causes for poor asthma management and asthma control.

Supplemental Material

Illness_representations_predicts_asthma_control_final_-_online_supplement_rev1 – Supplemental material for Leventhal’s common-sense model and asthma control: Do illness representations predict success of an asthma rehabilitation?

Supplemental material, Illness_representations_predicts_asthma_control_final_-_online_supplement_rev1 for Leventhal’s common-sense model and asthma control: Do illness representations predict success of an asthma rehabilitation? by Lea I Achstetter, Konrad Schultz, Hermann Faller and Michael Schuler in Journal of Health Psychology

Footnotes

Acknowledgements

The authors wish to thank Gunda Musekamp, Katja Spanier, Monika Schwarze, Inge Ehlebracht-König Christoph Gutenbrunner, Roland Kirchhof, Dragan Stojanovic, and Oliver Göhl.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by the German Federal Ministry of Education and Research (Bundesministerium fuer Bildung und Forschung) and supported by the Consortium to promote rehabilitation at the Clinic Bad Reichenhall (Arbeitsgemeinschaft zur Foerderung der Rehabilitation an der Klinik Bad Reichenhall e.V.).

Supplemental Material

Supplemental material for this article is available online.

References

Asparouhov

Muthen

(2010) Chi-square statistics with multiple imputation. Available at: https://www.statmodel.com/download/MI7.pdf

Baines

Wittkowski

(2013) A systematic review of the literature exploring illness perceptions in mental health utilising the self-regulation model. Journal of Clinical Psychology in Medical Settings 20(3): 263–274.

Bollen

(1989) Structural Equations with Latent Variables. Oxford: John Wiley & Sons.

Boulet

FitzGerald

Reddel

(2015) The revised 2014 GINA strategy report: Opportunities for change. Current Opinion in Pulmonary Medicine 21(1): 1–7.

Burgess

McCaffery

Powell

. (2015) The influence of asthma control on psychosocial outcomes for pregnant women with asthma. Journal of Asthma 52: 1013–1019.

Byer

Myers

(2000) Psychological correlates of adherence to medication in asthma. Psychology, Health & Medicine 5(4): 389–393.

Calfee

Katz

Yelin

. (2006) The influence of perceived control of asthma on health outcomes. Chest 130(5): 1312–1318.

Demoly

Paggiaro

Plaza

. (2009) Prevalence of asthma control among adults in France, Germany, Italy, Spain and the UK. European Respiratory Review 18(112): 105–112.

Diefenbach

Leventhal

(1996) The common-sense model of illness representation: Theoretical and practical considerations. Journal of Social Distress and the Homeless 5(1): 11–38.

10.

Glattacker

Bengel

Jäckel

(2009) Die deutschsprachige Version des Illness Perception Questionnaire-Revised. Psychometrische Evaluation an Patienten mit chronisch somatischen Erkrankungen. Zeitschrift für Gesundheitspsychologie 17(4): 158–169.

11.

Glattacker

Heyduck

Meffert

(2013a) Illness beliefs and treatment beliefs as predictors of short-term and medium-term outcome in chronic back pain. Journal of Rehabilitation Medicine 45(3): 268–276.

12.

Glattacker

Heyduck

Meffert

(2013b) Illness beliefs and treatment beliefs as predictors of short and middle term outcome in depression. Journal of Health Psychology 18(1): 139–152.

13.

Glattacker

Opitz

Jäckel

(2010) Illness representations in women with fibromyalgia. British Journal of Health Psychology 15(Pt 2): 367–387.

14.

Global Initiative for Asthma (GINA) (2006) Global Strategy for Asthma Management and Prevention (revised 2006). GINA. Available at: www.ginasthma.org.

15.

Global Initiative for Asthma (GINA) (2014) Global Strategy for Asthma Management and Prevention (revised 2014). GINA. Available at: www.ginasthma.org.

16.

Hamaker

Kuiper

Grasman

RPPP

(2015) A critique of the cross-lagged panel model (Empirical Study Mathematical Model). Psychological Methods 20(1): 102–116.

17.

Heyduck

Bengel

Farin-Glattacker

. (2015) Adolescent and parental perceptions about asthma and asthma management: A dyadic qualitative analysis. Child: Care, Health and Development 41: 1227–1237.

18.

Horne

Weinman

(2002) Self-regulation and self-management in asthma: Exploring the role of illness perceptions and treatment beliefs in explaining non-adherence to preventer medication. Psychology & Health 17(1): 17–32.

19.

IBM Corp. (2014) IBM SPSS Statistics for Windows, version 23.0. Armonk, NY: IBM Corp.

20.

Jessop

Rutter

(2003) Adherence to asthma medication: The role of illness representations. Psychology & Health 18(5): 595–612.

21.

Jia

Zhang

. (2013) The asthma control test and asthma control questionnaire for assessing asthma control: Systematic review and meta-analysis. Journal of Allergy and Clinical Immunology 131(3): 695–703.

22.

Kaptein

Hughes

Scharloo

. (2008) Illness perceptions about asthma are determinants of outcome. Journal of Asthma 45(6): 459–464.

23.

Kaptein

Klok

Moss-Morris

. (2010) Illness perceptions: Impact on self-management and control in asthma. Current Opinion in Allergy and Clinical Immunology 10(3): 194–199.

24.

Kazis

Anderson

Meenan

(1989) Effect sizes for interpreting changes in health status. Medical Care 27(3 Suppl.): S178–S189.

25.

Leventhal

Bodnar-Deren

Breland

. (2012) Modeling health and illness behavior: The approach of the commonsense model. In: Baum

Revenson

Singer

(eds) Handbook of Health Psychology (2nd edn). New York: Psychology Press, pp. 3–35.

26.

Lingner

Ernst

Großhennig

. (2015) Asthma control and health-related quality of life one year after inpatient pulmonary rehabilitation: The ProKAR Study. Journal of Asthma 52: 614–621.

27.

Lim

. (2014) Psychiatric comorbidities in Asian adolescent asthma patients and the contributions of neuroticism and perceived stress. Journal of Adolescent Health 55(2): 267–275.

28.

Meredith

(1993) Measurement invariance, factor analysis and factorial invariance. Psychometrika 58(4): 525–543.

29.

Moss-Morris

Weinman

Petrie

. (2002) The revised illness perception questionnaire (IPQ-R). Psychology & Health 17(1): 1–16.

30.

Muthén

(2015) Mplus User’s Guide. Los Angeles, CA: Muthén & Muthén.

31.

Oatley

Johnson-Laird

(2014) Cognitive approaches to emotions. Trends in Cognitive Sciences 18(3): 134–140.

32.

Pedersen

(2010) From asthma severity to control: A shift in clinical practice. Primary Care Respiratory Journal 19(1): 3–9.

33.

Powell

McCaffery

Murphy

. (2011) Psychosocial outcomes are related to asthma control and quality of life in pregnant women with asthma. Journal of Asthma 48(10): 1032–1040.

34.

Schatz

Sorkness

. (2006) Asthma control test: Reliability, validity, and responsiveness in patients not previously followed by asthma specialists. Journal of Allergy and Clinical Immunology 117(3): 549–556.

35.

Schuler

Faller

Wittmann

. (2016) Asthma control test and asthma control questionnaire: Factorial validity, reliability and correspondence in assessing status and change in asthma control. Journal of Asthma 53: 438–445.

36.

Schuler

Musekamp

Bengel

. (2014) Measurement of stable changes of self-management skills after rehabilitation: A latent state-trait analysis of the Health Education Impact Questionnaire (heiQ). Quality of Life Research 23(9): 2531–2543.

37.

Schuler

Musekamp

Faller

. (2013) Assessment of proximal outcomes of self-management programs: Translation and psychometric evaluation of a German version of the Health Education Impact Questionnaire (heiQ). Quality of Life Research 22(6): 1391–1403.

38.

Searle

Norman

Thompson

. (2007) Illness representations among patients with type 2 diabetes and their partners: Relationships with self-management behaviors. Journal of Psychosomatic Research 63(2): 175–184.

39.

Selig

Little

(2012) Autoregressive and cross-lagged panel analysis for longitudinal data. In: Laursen

Little

Card

(eds) Handbook of Developmental Research Methods. New York: Guilford Press, pp. 265–278.

40.

Sofianou

Martynenko

Wolf

. (2013) Asthma beliefs are associated with medication adherence in older asthmatics. Journal of General Internal Medicine 28(1): 67–73.

41.

Tiggelman

van

Ven

van Schayck

. (2014) The common sense model in early adolescents with asthma: Longitudinal relations between illness perceptions, asthma control and emotional problems mediated by coping. Journal of Psychosomatic Research 77(4): 309–315.

42.

Van Buuren

(2012) Flexible Imputation of Missing Data. Boca Raton, FL: CRC Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB