Modification and Psychometric Testing of the German-Language Revised Illness Perception Questionnaire (IPQ-R) in Occupational Dermatological Rehabilitation

Abstract

Purpose: This study aims at the modification and psychometric evaluation of the “revised Illness Perception Questionnaire” (IPQ-R) for occupational dermatological rehabilitation. Methods: First, the questionnaire was modified for application in occupational dermatology. Subsequently, 254 patients of an inpatient rehabilitation programme participated in a cross-sectional survey. Afterwards, the dimensional analysis of the IPQ-R was conducted using principal component analysis. Separate analyses were conducted for the illness representations and the causal attribution scale. Results: A total of 228 participants were included in the analysis (age: M = 48.2 years; SD = 12.0; 53.9% female). The patient acceptance of the questionnaire was high (response rate 87.3%; rate of completion between 92.5% and 98.4%, N = 254). The IPQ-R for occupational dermatology consists of 29 items in the domain of illness representations, which include seven factors (illness coherence, emotional representations, consequences: implications for the structuring of own life, consequences: financial and social impacts, treatment control, personal control, and timeline acute/chronic). Six of these scales have acceptable-to-good internal consistency (Cronbach’s α 0.72–0.84); for one scale, the internal consistency is Cronbach’s α = 0.66. A separate analysis of the causes resulted in eight factors (psychological causes at work and during leisure time, attributions outside the workplace, skin cleansing and skin protection measures, behaviour-related risk factors, causes at work, other risk factors, external factors that cannot be influenced by the person, and climatic influences) with a total of 30 items. Five of the eight scales have an acceptable-to-good internal consistency (Cronbach’s α 0.71–0.83), and three scales are just below the acceptable range (Cronbach’s α 0.63–0.66). Conclusion: Overall, the initial psychometric results of the IPQ-R for occupational dermatology were satisfactory. However, additional validation steps are still required. The following differences to the original model should be considered when interpreting the available results: the factor “timeline cyclical” could not be replicated in this field of application. Additionally, two factors with different thematic emphases in the “consequences” section, besides effects on the personal way of life, social and financial consequences, became visible as well.

1. Introduction

When a person is diagnosed with an illness, suffers an injury or notices signs of an illness, this typically leads to the development of personal ideas about the perceived health threat [1]. These so-called illness perceptions or illness representations, about the nature or characteristics of an illness, its development and course, as well as the possibilities of treatment, have a significant influence on the coping chosen by the individual [2].

1.1. Common Sense Model of Self-Regulation

The “Common Sense Model of Self-Regulation of Health and Illness” (CSM) according to Leventhal et al. [1] provides a theoretical framework to explain these self-regulation processes that take place on a cognitive and emotional level. The basic assumption of the model is that people endeavour to counteract a health threat through regulatory processes. Chosen coping strategies and their evaluation of success continuously influence the individual’s perceptions, which are therefore subject to dynamic change [3].

The CSM uses six dimensions to map assumptions about a disease [2,4]:

• Identity (e.g., symptoms causally associated and perceived with the disease, name of the disease);
• Cause (e.g., suspected causal links via the triggers of the patient’s own illness);
• Timeline (e.g., expectations about the duration and (acute/chronic or cyclical) course of the disease);
• Controllability (e.g., assumptions about the ability to influence the disease through one’s own behaviour or therapeutic interventions);
• Consequences (e.g., assumptions about the short and long term, psychological, social, emotional, physical and/or economic consequences of the disease);
• Illness coherence (e.g., extent to which the disease is comprehensible).

The influence of illness perceptions on various relevant outcomes (e.g., quality of life, depression) and on the choice of coping strategies (e.g., expressing feelings, problem-orientated behaviour, use of social support) has already been investigated for various diseases [5]. So, this led, amongst other things, the dimension “Illness coherence” has been integrated in the CSM over time [6]. With regard to eczematous skin diseases, Rocholl et al. [6] were able to show in their systematic review that these patients have complex illness perceptions, which were analysed using various qualitative and quantitative survey methods. Among other things, psychological and work-related consequences of the disease are described, e.g., fear of job loss. In addition, the skin disease is perceived as a major emotional burden. Wittkowski et al. [7] used the IPQ-R with individuals suffering from atopic dermatitis and reported that stress, heredity, and emotional well-being were named as the most frequent causes. In addition, more than 75% of participants stated that they perceived their condition as chronic. The results of this study indicate that the illness perceptions and emotional reactions are only partially related to the severity of the disease [7]. For patients with psoriasis, Solmaz et al. [8] reported that people’s quality of life is positively influenced when they understand their disease and consider it controllable. A negative influence on quality of life was found in people with multiple symptoms and higher disease severity, who experienced impairments due to the disease and who highlighted personal causes of the disease (e.g., own behaviour, negative attitudes).

1.2. Instrument: Illness Perception Questionnaire—Revised Version

The revised version of the IPQ-R [9], which is available in various translations and has been validated for different clinical pictures, enables the systematic investigation of illness perceptions. While it has already been adapted and validated for individual skin diseases (e.g., psoriasis [10,11,12], atopic dermatitis [7,13]), no psychometrically tested version exists for patients with work-related hand eczema. There is also a short version of the IPQ-R: “Brief Illness Perception Questionnaire” (B-IPQ). This is recommended, particularly due to its significantly shorter length, if the instrument is to be used several times within a short period of time. In this way, the burden on participants completing the questionnaires can be minimised [14].

The IPQ-R is an instrument for systematically assessing illness perceptions [9] and consists of three sections. In the first section, patients’ theories about the symptoms associated with their illness are recorded. This so-called “identity scale” is regarded as a sum score and is therefore not considered below when determining the present factor structure [13,15]. The second section consists of seven subscales: timeline acute/chronic, consequences, personal control, treatment control, coherence, cyclical timeline, and emotional representations. The original English-language scale consists of 38 items in this section [9]. The German version by Gaab et al. [16] used in this study consists of 32 items, which can also be assigned to these seven factors [16,17]. In the course of translating the IPQ-R into German, the factorial structure was largely confirmed, although six items were excluded due to insufficient factor loading [17]. Based on their results, Glattacker et al. [17] concluded that illness perceptions “can be measured relatively reliably and validly with the German-language IPQ-R” (own translation; p. 167). The items are rated on a five-point Likert scale (“strongly disagree” to “strongly agree”). High values indicate more pronounced beliefs in the area of the respective scales: a high value in the area of “consequences”, for example, suggests more pronounced assumptions about consequences, whereas a high value on the “coherence” scale implies a better understanding of the illness. The last section of the IPQ-R (18 items) deals with possible causes, which are also assessed using the five-point Likert scale. In the past, the IPQ-R has shown good psychometric properties in terms of validity and reliability for various diseases and indication-specific fields of application [9,13,16,18].

1.3. Application Area: Occupational Dermatology

Hand eczema (e.g., irritant, allergic, or atopic hand eczema, as well as mixed forms that have both occupational and congenital components) has been one of the most frequently reported occupational diseases in Germany for decades. The prevalence and incidence are high due to the pronounced skin exposure to irritants and allergens in a wide variety of occupational groups, e.g., health-related, domestic, or metalworking professions [19]. Outpatient and inpatient measures for individual prevention have therefore been implemented in standard pathways of patient care provided by the statutory social accident insurance in Germany. Measures of individual prevention in the administrative “dermatologist’s procedure” [20] of the statutory social accident insurance are understood to be graduated interventions adapted to the individual severity of the illness. This includes both preventive programmes in the outpatient sector and a so-called “inpatient treatment procedure”, which corresponds to an inpatient rehabilitation programme [21]. Clinically severe occupational dermatoses that are resistant to outpatient treatment are treated in specialised clinics where, in addition to dermatological diagnostics and therapy, behavioural interventions (e.g., patient education and counselling) are also carried out [21]. In this context, mixed and overlapping aetiologies of hand eczema are very common, frequently involving atopic hand eczema. In the case of work-related hand eczema in particular, studies indicate that individual patient behaviour (e.g., application of skin protection and skin care measures as well as adherence to therapy) can make a significant contribution to ensuring that patients can continue to work in the long term, stabilise the skin condition, and improve their quality of life [22]. Patient education and counselling are therefore key elements that can lead to supporting corresponding behavioural changes in patients [22].

The database resulting from the use of the “Illness Perception Questionnaire—revised version” (IPQ-R) enables a detailed characterisation of the rehabilitation cohort. Knowledge of illness perceptions offers those providing treatment the opportunity to understand observed coping behaviour and adherence. In addition, they can be the starting point for interventions that lead to a change in health or coping behaviour by processing existing illness perceptions [2]. This is particularly important for patients with chronic illnesses [1].

To date, there is no suitable instrument for recording the specific disease assumptions of this group of people, e.g., in order to address them specifically in patient education and counselling. However, the targeted recording and description of illness perceptions in a professional context is a necessary prerequisite. This justifies the need to adapt the IPQ-R.

1.4. Objective

The aims of the present study are:

(1)

to develop a German-language version of the IPQ-R adapted for occupational dermatology based on the existing instrument [9] and;

(2)

to use this modified instrument in occupational dermatological rehabilitation for the first time and to test it psychometrically.

2. Material and Methods

2.1. Modifications and Item Development

The authors of the IPQ-R recommend adapting the generically developed questionnaire to the respective research context [9]. As a basis for the adaptations, versions of the IPQ-R adapted to specific dermatological disease patterns were initially reviewed. Among other things, it was found that the IPQ-R was used unchanged for psoriasis patients, e.g., in Solmaz et al. [8]. A disease-related adaptation of the Illness Perception Questionnaire (IPQ) for this group of people was carried out, for example, by Scharloo et al. [10] of the possible causal attributions: the causal scale was shortened to “stress”, as this was considered a particularly important causal attribution [10]. In the study by Wittkowski et al. [13], the illness perceptions in patients with atopic dermatitis were analysed using the IPQ-R. The following changes were made to the instrument: “Itching” was included as a symptom in the identity scale. “My illness” was replaced by “My eczema” [13]. In addition, survey instruments for assessing skin condition (e.g., “Nordic Occupational Skin Questionnaire” [NOSQ] [23], “Osnabrück Hand Eczema Severity Index” [OHSI] [24]) were analysed to identify dermatologically relevant symptoms and signs of illness. An instrument that considers both skin protection measures and climatic influences could not be identified. The following modifications were therefore made for the survey reported here: rephrasing the term “my illness” is one of the most common adaptations of the IPQ-R (e.g., “my weight problem” [15]) [16]. Therefore, the term “disease” was replaced by the term “hand eczema” (see

Table 1. Presentation of the modified items of the IPQ-R (changes are highlighted in bold and italics).

IP1 *	My hand eczema will last a short time
IP2	My hand eczema will last a long time
IP3 *	My hand eczema will pass quickly
IP4	I expect to have my hand eczema for the rest of my life
IP5	My hand eczema has major consequences on my life
IP6 *	My hand eczema does not have much effect on my life
IP7	My hand eczema strongly affects the way others see me
IP8	My hand eczema has serious financial consequences
IP9	My hand eczema causes difficulties for those who are close to me
IP10	There is a lot which I can do to control my symptoms
IP11	What I do can determine whether my hand eczema gets better or worse
IP12	The course of my hand eczema depends on me
IP13	I have the power to influence my hand eczema
IP14 *	My hand eczema will improve in time
IP15 *	There is very little that can be done to improve my hand eczema
IP16	My treatment will be effective in curing my hand eczema
IP17	The negative effects of my hand eczema can be prevented (avoided) by my treatment
IP18	My treatment can control my hand eczema
IP19 *	The symptoms of my hand eczema are puzzling to me
IP20 *	My hand eczema is a mystery to me
IP21 *	I don’t understand my hand eczema
IP22 *	My hand eczema doesn’t make any sense to me
IP23	I have a clear picture or understanding of my condition
IP24	The symptoms of my hand eczema a great deal from day to day
IP25	My symptoms come and go in cycles
IP26	My hand eczema is very unpredictable
IP27	I go through cycles in which my hand eczema gets better and worse
IP28	I get depressed when I think about my hand eczema
IP29	When I think about my hand eczema I get upset
IP30	My hand eczema makes me feel angry
IP31	Having this hand eczema makes me feel anxious
IP32	My hand eczema makes me feel afraid

* = negatively polarised items.

).

Secondly, based on content considerations, four new items were formulated for the section on the assessment of illness perceptions in the area of consequence expectations (see

Table 2. Supplementary formulated items for recording illness perceptions.

N1	My hand eczema has an impact on my professional activity
N2	My hand eczema has an impact on activities of daily life (e.g., cleaning, washing dishes, etc.)
N3	My hand eczema has an impact on my leisure activities (e.g., hobbies, social activities, etc.)
N4	My hand eczema has an impact on my partnership

). These additions resulted from a patient survey in health psychology patient seminars during occupational dermatological rehabilitation and based on the results of Rocholl et al. [6]. From this, the research group developed an item pool in the format of an expert discussion. A first version of the resulting modified instrument was used as part of a student qualification project [25]. The items were tested with four patients using the “thinking aloud” method [26]. The instrument was then revised considering the results of the qualification work in a further expert discussion. The final questionnaire used in the study context was trialled in a pre-test with seven patients.

The 18 original items of the “Causes” scale in Section 3 of the German translation were used unchanged, as these causes (e.g., diet) can be categorised as patient-relevant based on the feedback from patients in skin protection patient education courses. This scale was supplemented by 15 additional, potentially relevant causes from the above-mentioned item pool, which were derived from dermatological survey instruments and, among other things, establish a link to the occupational activity, as well as the results of Rocholl et al. [6] with pivotal aspects relevant to the field of application of work-related skin diseases (see

Table 3. Supplementary formulated disease-specific causes.

NC1	Hormone fluctuations
NC2	Activities at work
NC3	Activities outside of work
NC4	Skin-irritating substances at work
NC5	Skin-irritating substances outside of work
NC6	Sensitising substances at work
NC7	Sensitising substances outside of work
NC8	Hand washing
NC9	Hand disinfection
NC10	Creams
NC11	Wearing gloves
NC12	Inadequate/lack of skin protection at work
NC13	Climatic influences at work (e.g., cold, heat, dust exposure, solar radiation)
NC14	Climatic influences outside of work (e.g., cold, heat, dust exposure, solar radiation)
NC15	Working atmosphere (e.g., cooperation with colleagues/employer)

).

2.2. Data Collection and Sample

The IPQ-R was used as part of a cross-sectional study at the Institute for Interdisciplinary Dermatological Prevention and Rehabilitation (iDerm, Osnabrück, Germany) between June 2020 and May 2021. The survey sample comprised rehabilitants who were admitted to an inpatient rehabilitation programme due to an occupational skin disease [19]. Participants were recruited consecutively as part of the “welcome seminar”, which all patients regularly attend on the day of arrival, and which focuses on organisational content. The data collection took place after the welcome seminar and therefore before the health education training and counselling sessions, in which topics regarding the pathogenesis and prevention of work-related skin diseases are discussed with the participants.

In addition to diagnosed hand eczema, the following inclusion criteria were defined: signed declaration of consent to participate in the study, age of majority, and sufficient German language skills to complete the questionnaire. The study information (e.g., on voluntariness and data protection) and recruitment were carried out on site, mainly by a principal investigator (MR). All respondents gave their written consent after receiving verbal and written information.

The project “Mixed-methods study to assess illness perceptions of patients with occupational contact dermatitis of the hands for enhancing patient education and counselling” (SubjeKt), in the context of which the cross-sectional survey was conducted, was reviewed and approved by the Ethics Committee of the Osnabrück University (vote 13/2020, 8 April 2020).

2.3. Statistical Data Analysis

A total of 34.7% (N = 88) of all data sets were randomly checked for data entry errors when transferring the completed questionnaire to SPSS Version 26 [27] in order to statistically ensure fewer than 1.5% errors. A fully checked data set contained all 246 items collected as part of the above-mentioned project. The random data check resulted in an average error rate of 0.6% per questionnaire. The incorrect entries were corrected in the course of the data check. The data analysis was carried out using the statistical programme IBM SPSS Version 26 [27]. Before the evaluation began, the missing values were analysed, and the plausibility of the data was checked. The data gaps were marginal; therefore, no formal test was conducted. Twenty-six incomplete data sets were excluded (missing values in the IPQ-R range of ≥5% [28]). In order to avoid bias, the missing values of data sets with <5% missing values in the IPQ-R range were imputed using the EM algorithm [29] implemented in SPSS Version 26 [27] for all items of the questionnaire and the personal characteristics.

These analyses, as described below, constitute a first exploratory step: the structure was tested using an exploratory factor analysis (EFA) with the aid of principal component analysis and VARIMAX rotation in order to obtain a simple structure because this is the approach applied in the original IPQ-R validation [9]. Although methods that model common factors (e.g., principal axis factoring) are generally preferred for latent variable modelling, PCA provides a data driven description of variance and is appropriate for an initial exploratory investigation. Analogous to the procedure in the validation study of the original IPQ-R [9], the disease assumptions were analysed in a first step. In a second step, a separate EFA was implemented for the 33 causal attributions in order to determine the best fit [13,15]. The suitability of the variables was tested in advance using the Kaiser–Meyer–Olkin criterion (KMO > 0.50) and the Bartlett test for sphericity. The Kaiser-Guttman criterion (eigenvalue > 1) was used to identify the number of factors. Cronbach’s α (α > 0.70: acceptable; α > 0.80: good [30]) was calculated to analyse the internal consistency of the IPQ-R scales. The quality of the items was assessed on the basis of item difficulty (item means), the factorial loading structure (loadings > 0.50), and the corrected discriminatory power (>0.40 [31]). If the factor loading of >0.50 could not be achieved, the item was eliminated, provided that there were no content-related considerations preventing exclusion. If the factor loading of an item was divided into several factors, the allocation to a factor was based on content criteria. In addition, Pearson’s correlation coefficients were analysed to determine the intercorrelations between the scales of the questionnaire. The following correlations were assumed: A correlation can be expected between consequences and emotional representations [17,19]. In addition, correlations between the “Personal Control” and “Treatment Control” scales were assumed [19], as individual behaviour (e.g., treatment adherence) and treatment options are related in the case of work-related hand eczema. Due to the very heterogeneous results on the factors in the area of disease causes and the simultaneous lack of comparability due to the modifications made, no preliminary assumptions were formulated at this point.

3. Results

3.1. Sample Description and Acceptance of the Items

Of the 291 patients who attended the welcome seminar, 254 people took part in the survey (response rate: 87.3%). Due to the case exclusion described above, a total of 228 data sets were included in the analysis. The average age of the respondents was 48 years (M = 48.2; SD = 12.0). A detailed sample description is shown in

Table 4. Sample description of the non-imputed data set (N = 228).

	Frequency	Percent
Sex
Female	123	53.9%
Male	105	46.1%
Diverse	0	0.0%
Age group
18–20 years	1	0.4%
20–29 years	29	12.7%
30–39 years	24	10.5%
40–49 years	33	14.5%
50–59 years	108	47.4%
60–69 years	33	14.5%
Living in a partnership
Yes	171	75.0%
No	55	24.1%
Not specified	2	0.9%
Educational qualification
No school-leaving qualification	1	0.4%
Secondary school-leaving certificate	64	28.1%
Secondary school-leaving certificate	100	43.9%
Advanced technical college entrance qualification	30	13.2%
General higher education entrance qualification/abitur	19	8.3%
Bachelor’s degree	3	1.3%
Master/diploma	4	1.8%
Promotion	2	0.9%
Not specified	5	2.2%
Duration of the skin disease
≤1 year	25	11.0%
1 ≤ 2 years	54	23.7%
2 ≤ 5 years	56	24.6%
5 ≤ 10 years	35	15.4%
10 ≤ 15 years	17	7.5%
15 ≤ 20 years	15	6.5%
20 ≤ 30 years	7	3.1%
≥30 years	8	3.5%
Not specified	11	4.8%
Severity of hand eczema at the time of the survey (patient self-assessment)
Clear	3	1.3%
Almost clear	11	4.8%
Mild	26	11.4%
Moderate	131	57.5%
Severe	45	19.7%
Not specified/prefer not to say	12	5.2%

.

The completion rate, considering all participants without exclusion, as an indicator of the acceptance of the items (on the non-imputed data set, N = 254), was between 92.5% (IP3 * “My hand eczema will pass quickly”) and 98.4% (NC2 “Activities at work”, NC4 “Skin-irritating substances at work”, and NC8 “Hand washing”).

3.2. Factorial Structure of the Disease Assumptions

• Examination of the requirements and initial analysis

Before performing the principal component analysis, its requirements were checked. Both Bartlett’s test for sphericity (Chi-square(465) = 2976.39, p < 0.001) and the KMO value of 0.78 indicated the suitability of the 36 items for a factor analysis. A principal component analysis, which contained all disease assumptions (IP1–IP32 and N1–N4) and VARIMAX rotation with an eigenvalue > 1, showed the following results in the first analysis step: nine factors were formed. Supplementary observation of the screeplot suggested a solution with four factors. In the following, the eigenvalue was used as the basis for interpreting the results, as the factors according to this criterion were similar to the original model.

• Item allocation

The items were assigned to the scales based on the factor loading. Items with a loading >0.50 on one factor and with loadings <0.40 on all other factors were retained (see

Table 5. Factor analysis (principal component analysis with VARIMAX rotation) for the scales of disease assumptions and internal consistency (Cronbach’s α) on the imputed data set as well as the item characteristics (N = 228).

	Factor (Variance Explanation)								Item Characteristics
Items	Factor 1 (18.65%)	Factor 2 (12.01%)	Factor 3 (8.27%)	Factor 4 (6.82%)	Factor 5 (5.51%)	Factor 6 (5.33%)	Factor 7 (3.99%)	Factor 8 (3.37)	Communality	Arithmetic Mean	Corrected Item Discriminatory Power
1 Illness coherence (α = 0.84)
IP21 * I don’t understand my hand eczema	0.878	−0.131	−0.073	−0.046	0.026	0.089	0.076	−0.063	0.81	2.88 (1.16)	0.81
IP20 * My hand eczema is a mystery to me.	0.873	−0.122	−0.042	−0.041	−0.061	0.105	0.037	−0.047	0.80	2.88 (1.26)	0.79
IP19 * The symptoms of my hand eczema are puzzling to me	0.778	−0.015	0.111	−0.209	−0.035	0.149	−0.005	−0.003	0.69	2.79 (1.25)	0.67
IP22 * My hand eczema doesn’t make any sense to me	0.733	−0.067	0.001	−0.016	0.129	0.060	−0.021	−0.178	0.59	3.02 (1.14)	0.60
IP23 I have a clear picture or understanding of my condition	0.533	−0.174	0.140	0.035	0.108	0.019	0.025	0.393	0.50	3.18 (1.07)	0.40
2 Emotional representations (α = 0.82)
IP28 I get depressed when I think about my hand eczema	−0.116	0.777	0.150	0.173	−0.090	0.005	0.000	0.009	0.68	3.59 (1.11)	0.67
IP30 My hand eczema makes me feel angry	−0.117	0.756	0.238	0.086	−0.018	−0.014	0.029	0.069	0.66	3.29 (1.17)	0.63
IP31 Having this hand eczema makes me feel anxious	−0.162	0.724	−0.082	0.298	−0.106	−0.050	−0.026	0.271	0.73	3.21 (1.12)	0.69
IP32 When I think about my hand eczema I get upset	−0.185	0.710	−0.022	0.319	−0.023	−0.053	−0.057	0.292	0.73	3.11 (1.17)	0.70
IP29 It worries me when I think about my hand eczema	0.007	0.619	0.208	0.021	0.114	−0.100	0.140	−0.136	0.49	3.22 (1.28)	0.43
3 Consequences: implications for the structuring of own life (α = 0.78)
N2 My hand eczema affects the activities of daily living (e.g., cleaning, washing up, etc.)	0.067	0.184	0.840	0.100	0.071	0.008	0.111	0.037	0.77	4.33 (0.79)	0.74
N1 My hand eczema has an impact on my professional activity	0.188	0.042	0.745	0.097	0.104	−0.105	0.059	0.152	0.65	4.34 (0.75)	0.55
IP5 My hand eczema has major consequences on my life	−0.025	0.265	0.695	0.231	−0.057	0.006	0.137	0.045	0.63	4.09 (0.94)	0.65
N3 My hand eczema has an impact on my leisure activities (e.g., hobbies, social activities, etc.)	−0.112	0.039	0.693	0.382	−0.008	0.052	0.027	0.023	0.64	3.95 (1.02)	0.57
IP6 * My hand eczema does not have much effect on my life	−0.089	0.049	0.466	0.118	−0.079	0.130	0.174	−0.205	0.34	4.05 (1.11)	0.38
4 Consequences: financial and social impacts (α = 0.72)
IP9 My hand eczema causes difficulties for those who are close to me	−0.063	0.228	0.198	0.710	−0.098	−0.059	0.028	0.091	0.62	3.17 (1.21)	0.58
IP7 My hand eczema strongly affects the way others see me	−0.109	0.223	0.100	0.706	−0.004	0.065	−0.029	0.008	0.58	2.91 (1.18)	0.55
IP8 My hand eczema has serious financial consequences	0.090	0.109	0.243	0.684	0.034	0.060	−0.096	−0.084	0.57	2.77 (1.18)	0.46
N4 My hand eczema has an impact on my partnership	−0.157	0.074	0.182	0.615	−0.138	−0.053	0.076	0.147	0.49	2.84 (1.26)	0.45
5 Treatment control (α = 0.81)
IP18 My treatment can control my hand eczema	0.058	−0.005	0.010	−0.087	0.845	0.211	0.087	0.017	0.78	3.55 (0.86)	0.70
IP17 The negative effects of my hand eczema can be prevented (avoided) by my treatment	0.140	0.025	−0.097	−0.041	0.840	0.198	0.026	−0.108	0.79	3.43 (0.92)	0.70
IP16 My treatment will be effective in curing my hand eczema	−0.086	−0.064	0.151	−0.057	0.784	−0.024	−0.151	0.022	0.68	3.54 (0.86)	0.57
6 Personal control (α = 0.75)
IP11 What I do can determine whether my hand eczema gets better or worse	0.129	−0.033	0.120	−0.055	0.043	0.810	0.098	0.024	0.70	3.25 (1.09)	0.60
IP12 The course of my hand eczema depends on me	0.078	0.096	0.120	−0.024	0.026	0.796	−0.171	0.009	0.70	2.80 (1.05)	0.55
IP13 I have the power to influence my hand eczema	0.185	−0.126	−0.115	0.120	0.207	0.707	−0.004	−0.028	0.62	2.82 (1.10)	0.59
IP10 There is a lot which I can do to control my symptoms	0.011	−0.233	−0.174	0.003	0.356	0.554	0.054	0.174	0.55	3.17 (1.04)	0.44
7 Timeline acute/chronic (α = 0.66)
IP3 * My hand eczema will pass quickly	0.104	−0.055	0.013	0.072	0.014	−0.001	0.784	0.105	0.65	4.04 (0.85)	0.47
IP1 * My hand eczema will last a short time	−0.057	−0.004	0.218	−0.099	−0.077	0.117	0.755	−0.160	0.68	4.23 (0.81)	0.49
IP2 My hand eczema will last a long time	0.026	0.137	0.160	−0.018	0.022	−0.142	0.683	0.118	0.55	4.00 (0.85)	0.45
8 Timeline cyclical (α = 0.46)
IP24 The symptoms of my hand eczema change greatly from day to day	−0.080	0.182	−0.001	0.114	−0.066	0.110	0.010	0.744	0.62	3.28 (1.00)	0.29
IP26 My hand eczema is very unpredictable	−0.373	0.210	0.126	−0.011	0.008	−0.028	0.316	0.499	0.55	3.61 (0.98)	0.29

Notes: N = 228; items that load highest on the respective scale with at least with a factor loading of >0.5 in accordance with the hypothesis are highlighted. For items that load on several factors, the cross-loading is highlighted in bold. * Negatively polarised, recoded items; factors and items that had to be excluded in the following item and scale analysis are crossed out; cross-loading <0.40 item and scale analysis of the disease assumptions.

). After an initial factor analysis, item IP25 (“My symptoms come and go in cycles.”) was excluded due to its low factor loading (<0.50) and existing cross-loadings. When interpreting the content of this first structure, it was already noticeable that one factor contained items from two scales (IP15 * “There is very little that can be done to improve my hand eczema”; treatment control and IP25 “My symptoms come and go in cycles”; timeline cyclical). Item IP15 * was initially retained. Furthermore, after the next analysis step, item IP27 (“I go through cycles in which my hand eczema gets better and worse”) was eliminated for the same reasons (factor loading <0.5) as item IP25. In addition, content analysis of this second structure showed that the items IP4 (“I expect to have my hand eczema for the rest of my life”) and IP14 * (“My hand eczema will improve in time”; both time course) and IP15 * (treatment control; “There is very little that can be done to improve my hand eczema.”) loaded on one factor. Finally, three further items were excluded due to insufficient factor loadings (IP4, IP14 *) and due to ambiguous loadings over time and a resulting ambiguous assignment (IP15 *).

• Final analysis results

The EFA with the remaining 31 items resulted in eight factors (eigenvalue > 1), which explain 63.95% of the total variance and which can be interpreted meaningfully in terms of content based on the original German model. Of these eight factors, three factors correspond to this original model (“Coherence”, “Emotional Representations” and “Personal Control”). There are two deviations with regard to the other five factors: the consequences are divided between two factors in this structure. These factors consist of two or three original items (items with the suffix IP) and one or three items newly formulated for occupational dermatology (items with the suffix N; see Table 5). Items were removed from the factors “treatment control” (IP15 *), “timeline acute/chronic” (IP4 and IP14), and “timeline cyclical” (IP25 and IP27) due to the procedure described above. Overall, this model appears plausible in terms of content.

• Result after scale analysis

As the items IP24 “The symptoms of my hand eczema a great deal from day to day” and IP26 “My hand eczema is very unpredictable” were removed in the subsequent scale analysis due to low corrected discriminatory power (see Table 5), the resulting questionnaire consists of 29 items distributed across seven factors. Table 5 summarises selected characteristic values of the 29 scale items.

Six of the seven scales were considered have an acceptable-to-good internal consistency (Cronbach’s α 0.72–0.84), and one scale is marginally below the acceptable range (Cronbach’s α = 0.66; see Table 5). The corrected item-scale correlations of the eight-factor solution were considered as a measure of the discriminatory power of the items. These achieved values of >0.40 for 28 of the 31 items. Item IP6 * (“My hand eczema does not have much effect on my life”), whose corrected discriminatory power was just under 0.40, was left in the scale due to its significance in terms of content.

• Intercorrelations

Table 6. Intercorrelations of the modified scales to disease assumptions in patients with work-related hand eczema.

	1	2	3	4	5	6	7
1. Illness coherence	1	−0.278 **	−0.009	−0.166 *	0.112	0.251 **	0.052
2. Emotional representations		1	0.328 **	0.454 **	−0.085	−0.153 *	0.073
3. Consequences: implications for the structuring of own life			1	0.465 **	0.001	0.000	0.271 **
4. Consequences: financial and social impacts				1	−0.132 *	−0.033	0.023
5. Treatment control					1	0.332 **	−0.022
6. Personal control						1	−0.019
7. Timeline acute/chronic							1

* The correlation is significant at the 0.05 level (two-sided). ** The correlation is significant at the 0.01 level (two-sided).

shows the intercorrelations between the modified scales on disease assumptions with the highest correlation being between factors 3 and 4, addressing different consequence areas (r = 0.47). The second-highest correlation being between factors 2 and 4 (r = 0.45) corresponds to our prior assumption.

3.3. Factorial Structure of the Causes of Disease

• Examination of the requirements and initial analysis

Before conducting the principal component analysis, its requirements were checked. Both Bartlett’s test for sphericity (Chi-square(528) = 2727.50; p < 0.001) and the KMO value of 0.79 indicated the suitability of the 33 items for a factor analysis. A principal component analysis containing all 33 causes (C1–C18 and NC1–NC15), with VARIMAX rotation with an eigenvalue > 1, showed the following results: nine factors were formed. Supplementary observation of the screeplot suggested a solution with three factors. In the following, the eigenvalue was used as the basis for interpreting the results, as the factors were more differentiated in terms of content according to this criterion.

• Item allocation and item exclusion

The items were assigned to the scales based on the factor loading. Items with a loading >0.50 and with loadings <0.40 on all other factors were retained (see

Table 7. Factor analysis (principal component analysis with VARIMAX rotation) for the cause scales and internal consistency (Cronbach’s α) on the imputed data set as well as the item characteristics (N = 228).

	Factor (Variance Explanation)								Item Characteristics
Items	Factor 1 (21.73%)	Factor 2 (8.82%)	Factor 3 (6.72%)	Factor 4 (6.24%)	Factor 5 (5.76%)	Factor 6 (4.91%)	Factor 7 (4.15%)	Factor 8 (3.88%)	Communality	Arithmetic Mean	Corrected Item Discriminatory Power
1 Psychological causes at work and during leisure time (α = 0.83)
C12 My emotional state, e.g., feeling down, lonely, anxious, empty	0.718	0.062	0.123	0.217	0.036	0.189	0.068	0.099	0.63	2.50 (1.13)	0.67
C9 My attitude, e.g., thinking about life negatively	0.711	0.092	0.104	0.296	0.076	0.009	0.002	−0.049	0.62	2.16 (0.94)	0.63
C10 Family problems or worries caused my illness	0.710	0.127	0.170	0.130	−0.129	0.130	−0.013	0.078	0.61	2.25 (1.02)	0.63
C1 Stress and worry	0.710	0.107	0.093	−0.228	−0.059	0.180	0.089	0.120	0.63	3.54 (1.12)	0.55
NC15 Working atmosphere (e.g., cooperation with colleagues/employer)	0.591	−0.010	−0.068	0.141	0.189	−0.009	0.153	0.325	0.54	2.56 (1.17)	0.53
C17 My personality	0.565	0.128	0.091	0.327	0.058	−0.161	0.156	−0.051	0.51	2.03 (0.95)	0.49
C11 Overwork	0.558	0.067	0.158	−0.001	0.034	−0.046	0.315	0.090	0.45	2.88 (1.13)	0.52
2 Attributions outside the workplace (α = 0.79)
NC5 Skin-irritating substances outside of work	0.061	0.876	0.085	−0.008	0.055	−0.060	0.078	0.147	0.81	3.04 (1.13)	0.73
NC7 Sensitising substances outside of work	0.123	0.863	0.081	0.125	0.081	0.083	0.098	0.063	0.81	2.95 (1.10)	0.73
NC3 Activities outside work	0.198	0.677	0.226	−0.004	0.045	−0.020	−0.023	0.137	0.57	2.99 (1.08)	0.57
NC10 Creams	0.117	0.429	0.416	0.149	−0.015	0.176	0.175	−0.083	0.46	2.77 (1.12)	0.38
3 Skin cleansing and skin protection measures (α = 0.71)
NC8 Hand washing	0.117	0.192	0.786	0.045	0.025	0.011	0.135	0.121	0.70	3.49 (1.08)	0.58
NC9 Hand disinfection	0.146	0.106	0.754	−0.204	0.114	0.008	0.235	0.083	0.72	3.73 (1.10)	0.59
NC11 Wearing gloves	0.188	0.089	0.652	0.042	0.218	0.104	−0.108	−0.044	0.54	3.47 (1.12)	0.42
4 Behaviour-related risk factors (α = 0.64)
C16 Accident or injury	0.059	0.149	−0.060	0.726	0.135	0.107	0.035	−0.105	0.78	4.17 (0.78)	0.65
C14 Alcohol	0.261	0.007	0.014	0.717	0.005	0.011	0.181	0.162	0.66	3.77 (1.01)	0.50
C15 Smoking	0.212	−0.025	−0.013	0.630	−0.116	0.129	−0.010	0.226	0.65	4.19 (0.76)	0.49
5 Causes at work (α = 0.71)
NC4 Skin-irritating substances at work	0.031	0.065	0.068	0.056	0.855	−0.141	0.072	0.082	0.61	2.58 (1.04)	0.47
NC6 Sensitising substances at work	0.021	0.242	0.022	0.097	0.748	0.180	−0.004	−0.017	0.54	2.59 (1.10)	0.37
NC2 Activities at work	0.021	−0.130	0.225	−0.106	0.744	−0.042	−0.052	0.125	0.61	2.70 (0.94)	0.43
6 Other risk factors (α = 0.63)
C3 A germ or virus	0.191	0.033	−0.076	−0.073	0.032	0.742	0.069	0.072	0.61	2.58 (1.04)	0.47
C5 Chance or bad luck	−0.014	−0.107	0.134	0.167	−0.069	0.681	−0.009	−0.108	0.54	2.59 (1.10)	0.37
C4 Diet or eating habits	0.458	0.205	0.021	0.004	0.055	0.512	0.285	−0.110	0.61	2.70 (0.94)	0.43
C6 Poor medical care in my past	0.005	0.034	0.136	0.287	0.003	0.508	0.188	0.146	0.42	2.42 (1.00)	0.36
7 External factors that cannot be influenced by the person (α = 0.66)
C18 Altered immunity	0.129	0.164	0.150	0.092	0.010	0.086	0.741	−0.096	0.64	2.92 (1.11)	0.50
C13 Ageing	0.243	−0.275	0.066	0.192	0.019	0.055	0.639	0.378	0.73	2.51 (1.02)	0.46
NC1 Hormone fluctuations	0.362	0.130	0.077	0.103	−0.021	0.179	0.590	−0.176	0.58	2.39 (1.04)	0.46
C7 Pollution in the environment	−0.044	0.125	0.016	−0.083	0.006	0.453	0.507	0.274	0.56	2.90 (1.13)	0.37
8 Climatic influences (α = 0.76)
NC13 Climatic influences at work (e.g., cold, heat, dust exposure, solar radiation)	0.082	0.115	0.066	0.129	0.182	0.111	−0.008	0.844	0.80	3.10 (1.16)	0.62
NC14 Climatic influences outside of work (e.g., cold, heat, dust exposure, solar radiation)	0.258	0.392	0.080	0.078	−0.001	−0.068	−0.002	0.694	0.72	2.83 (1.12)	0.62

Notes: N = 228; items that load highest on the respective scale with at least with a factor loading of >0.5 are highlighted. For items that load on several factors, the cross-loading is highlighted in bold cross-loading <0.40; cross-loadings are shaded in grey in the table and are explained and justified in the “Further analysis” section.

). Based on content considerations, an eight-factor solution was specified in the next step. One item (C2 “Hereditary—it runs in my family”) could not achieve a factor loading of >0.4. It was excluded due to similarly high cross-loadings across three factors.

• Further analysis

As the interpretability of the model could not yet be improved by the exclusion, a further eight-factorial analysis was carried out without item C2. The items C8 “My own behaviour” and NC12 “Inadequate/lack of skin protection at work” achieved a factor loading of 0.40 and 0.41, respectively. They were excluded for this reason and due to similarly high cross-loadings across several factors.

A further analysis revealed the following: the item NC10 “Creams” achieved a factor loading of 0.43 (causes outside of work) and loaded with a factor loading of 0.42 on a second factor (skin cleansing and skin protection measures). As this cross-loading can be explained in terms of content and it is a significant item for the scale from a therapeutic perspective, the item was retained. Item C4 “Diet or eating habits” showed a factor loading of 0.51 and was retained despite a cross-loading of 0.46 due to the significance of the item from the patient’s perspective. In this model, C7 “Pollution in the environment” was also retained despite the existing cross-loading as, on the one hand, it can be meaningfully interpreted in terms of content, and, on the other hand, it is a significant item for the scale.

• Final analysis results

The resulting final model (KMO = 0.78; Bartlett test for sphericity: Chi-square (435) = 2491.07; p < 0.001) can be interpreted meaningfully in terms of content and is shown in Table 7. Items that are distributed over several factors (NC10, C4, C7) were checked with regard to their factor affiliation. The final model consists of eight factors and 30 items and explains 62.21% of the total variance. The scales are summarised below. A differentiated presentation can be found in Table 7. The scale “psychological causes at work and during leisure time” comprises seven items and covers possible causes that can be attributed to the mental state. The “Causes outside of work” scale, with a total of four items, asks about products and activities that could trigger skin changes outside of work. The “Skin cleansing and skin protection measures” scale uses three items to record hand hygiene measures and the use of protective gloves. The “Behaviour-related risk factors” scale uses three items to represent aspects that can influence skin health as a result of personal behaviour. The “Causes at work” scale uses three items to investigate products and activities that could trigger skin changes during work. The “Other risk factors” scale uses four items to record a wide variety of aspects that can affect skin health. The scale “External factors that cannot be influenced by the person” uses four items to ask about the effects of processes in which the immune system is involved. The “Climatic influences” scale represents possible environmental influences at work and during leisure time with two items. A high value for one of the cause scales describes a high attribution of cause by the patient in this area.

• Item and scale analysis of the causes of illness

Five of the eight scales have an acceptable-to-good internal consistency (Cronbach’s α 0.71–0.83), and three scales are marginally below the acceptable range (Cronbach’s α 0.63–0.66; see Table 7). The corrected item-scale correlations of the eight-factor solution were considered as a measure of the discriminatory power of the items. This reached values of >0.40 for 26 of the 30 items (see Table 7). The items whose corrected discriminatory power was just below 0.40 were retained as part of the scales due to the significance of the content.

• Intercorrelations

Table 8. Intercorrelations of the modified cause scales for patients with work-related hand eczema.

	1	2	3	4	5	6	7	8
1. Psychological causes	1	0.338 **	0.336 **	0.383 **	0.114	0.326 **	0.429 **	0.340 **
2. Causes outside of work		1	0.411 **	0.175 **	0.185 **	0.192 **	0.239 **	0.358 **
3. Skin cleansing and skin protection measures			1	0.037	0.268 **	0.170 *	0.274 **	0.219 **
4. Behavioural risk factors				1	0.055	0.264 **	0.258 **	0.224 **
5. Causes at work					1	0.035	0.058	0.195 **
6. Other risk factors						1	0.450 **	0.105
7. External factors that cannot be influenced by the person							1	0.214 **
8. Climatic influences								1

* The correlation is significant at the 0.05 level (two-sided). ** The correlation is significant at the 0.01 level (two-sided).

shows the intercorrelations between the modified scales for the causes with the highest correlation between factors 6 (other risk factors) and 7 (external factors that cannot be influenced by the person; r = 0.45). This seems plausible in view of the fact that item C3 “A germ or virus” in the original model [9] is an item of the factor “External factors that cannot be influenced by the person”. It is noticeable that almost all factors in the medium range correlate with the psychological causes in work and leisure time.

4. Discussion

In order to support a change in health and coping behaviour in patients and to enable the development of coherent disease theories, psychological and social factors must be considered and adequately addressed in interventions during a rehabilitation measure [2]. The present questionnaire can be used to investigate illness perceptions in individuals suffering from work-related hand eczema. The current analysis is the first to investigate the psychometric properties of the IPQ-R in patients undergoing occupational dermatological rehabilitation in order to enable the systematic recording of illness perceptions in this area in the future.

The factorial validity of the occupational dermatological version of the IPQ-R was tested using two exploratory factor analyses. The items on the disease assumptions were split up so that the causes could be analysed separately. An initially exploratory approach was chosen over a confirmatory approach, as this modification of the instrument was being tested for the first time.

4.1. Illness Perceptions

The factorial structure of the questionnaire was largely confirmed in this area. With the help of the IPQ-R adapted for occupational dermatology, the following dimensions of the disease assumptions can be mapped: “Illness coherence”, “Emotional representations”, “Consequences: implications for the structuring of own life “, “Consequences: Financial and social impacts”, “Treatment control”, “Personal control”, and “Timeline acute/chronic”. The dimension “timeline cyclical” is not represented. In this form of its first empirical test, the questionnaire shows satisfactory characteristic values for the disease assumptions. The psychometric test revealed good characteristic values at both item and scale level. Shortening the questionnaire by one dimension and individual items led to satisfactory internal consistency for most scales (Cronbach’s α 0.72–0.84). In the original study by Moss-Morris et al. [9], these values were between 0.79 and 0.89, and in a study with psoriasis patients [8], were between 0.63 and 0.86.

In the procedure orientated along the validation study [9] of assigning items with a factor loading >0.50 to a factor, the vast majority of items loaded according to the hypothesis. Of the original 36 items in the area of disease assumptions, five (IP4, IP14 *, IP15 *, IP25, and IP27) did not fulfil this criterion. These were removed from the item pool. Two further items (IP24 and IP26) were also removed due to insufficient corrected discriminatory power. Overall, the scale intercorrelations indicate a plausible structure of the disease assumptions in terms of content. It should be noted, however, that no objective criterion was applied. The factors are analysed in detail below: as in the original German model, the subscales “Illness coherence”, “Emotional representations” and “Personal control” could be mapped. For the dimensions “Treatment control”, “Timeline acute/chronic”, and “Consequences”, deviations from the original model resulting from the shortening of the instrument described above must be considered in the interpretation.

4.2. Treatment Control

This factor can be described without the item IP15 * (“There is very little that can be done to improve my hand eczema.”). It has already been described as a conspicuous item in other studies [13]. For example, IP15 * was found to be problematic by Glattacker et al. [17] when testing the German version of the IPQ-R. Here, it loaded ambiguously on several factors. This can possibly be explained by the fact that the content of the item could also be assigned to personal control [17]. In the future, a positive formulation of the item could be considered and trialled (e.g., “There is a lot that can be done about my hand eczema”) [15].

4.3. Timeline (Acute/Chronic)

In the IPQ-R version reported here, this factor consists of three instead of five items. Glattacker et al. [17] also found this factor to be ambiguous. Although all items on the time course loaded on one factor, two further items (IP5 and IP6 *) from the “consequences” dimension could also be assigned to this factor. Wittkowski et al. [13], on the other hand, were able to demonstrate the stability of the factor in the English-language version of the IPQ-R for people with atopic dermatitis. This factor should therefore be tested in further studies in terms of content and psychometrics. In our data, the scale’s internal consistency is marginally below the acceptable range (Cronbach’s α = 0.66; see Table 5), so we recommend that future studies re-examine this scale both content-wise and psychometrically.

4.4. Consequences

The items of the “Consequences” subscale are distributed across two factors, which correlate with each other to a correspondingly high degree. This is not least due to the addition of additional items in this instrument. This correlation seems plausible, as the organisation of one’s own life is related to social and financial consequences. The extent to which these factors prove to be stable should be examined in further studies. The correlations between the subscales “emotional representations” and “consequences” also appear plausible in terms of content and were already shown by Glattacker et al. [17] in the course of testing the German-language version: many of the consequences mentioned (e.g., IP5) include emotional aspects [15]. However, this could contribute to an overlap of the subscales. In Wittkowski et al. [13], a factor consisting of items from both dimensions was found in the EFA for people with atopic dermatitis. There was also a strong correlation (r = 0.77, p < 0.01) between these two factors [13]. This supports the assumption of a possible intersection.

4.5. Timeline (Cyclical)

The factor “timeline (cyclical)” could not be replicated. Two items (IP25 and IP27) were eliminated due to insufficient factor loadings with simultaneous cross-loadings; two further items (IP24 and IP26) were subsequently removed due to insufficient discriminatory power. In their analysis of this factor, Wittkowski et al. [13] also found cross-loadings for two of the four items included. This seems plausible in terms of content, as atopic dermatitis, which can also occur together with irritant (hand) eczema as a mixed form, for example, is less easy to identify “as a pattern” due to its multifactorial pathogenesis and variable course compared to exclusively occupational diagnoses or other diseases. This scale is therefore not initially part of the IPQ-R modified for occupational dermatology. The development of an alternative should be examined in future works.

4.6. Causes

In their systematic review, Rocholl et al. [6] were able to show that both endogenous and exogenous causes are reported by patients with eczematous skin diseases. By modifying the scales as described above, it was possible to achieve an optimised factor structure for occupational dermatology that takes both endogenous and exogenous causes into account. However, due to the newly added items in the area of disease causes, these scales are only comparable with the original structure, according to Moss-Morris et al. [9], to a limited extent. This should be checked for stability in further analyses, particularly with regard to the following anomalies:

 (I)

Behavioural risk factors, such as substance use (C14 and C15), diet, and eating habits (C4), are distributed across two factors (factors 4 and 6).

(II)

C5 “chance or bad luck” and C16 “accident or injury”, which form the construct “chance” in the original model, are distributed over two risk-associated factors in this model. The instability of the original factor is also evident in other studies [13,15]: Wittkowski et al. [13] were also unable to replicate the factors “chance” and “immunity” in their EFA.

Overall, the scale intercorrelations identify a plausible structure for the modified cause scales. It should be noted that no objective criterion was applied here either. The correlation of almost all factors with the factor “Psychological causes at work and during leisure time” appears to be explainable, as all causes can be significant at this level. The correlations are at most in the low-to-medium range, so that there appears to be no redundancy of content between the scales.

4.7. Strengths and Limitations

One strength of this work lies in the methodological adaptation of the IPQ-R: the added items are based, among other things, on the results of a systematic review [6]. To our knowledge, this is the first study to transfer the IPQ-R to the field of work-related skin diseases and to analyse it psychometrically. In addition, a high completion rate, understood as an indicator of patient acceptance of the items, and the response rate support the assumption that this version of the IPQ-R is suitable for use in research. Participation was voluntary, yet the majority of rehabilitants took part in the survey. However, a selection bias cannot be completely ruled out, which limits the generalisability of the study. Furthermore, no information on non-respondents (12.7%, N = 37) and their motives is available.

First, we emphasise that this work represents an initial psychometric evaluation, which therefore requires further research. With respect to the scale for investigated causes, the number of items results in a unfavourable item-to-sample-size ratio, which may affect factor stability. However, since patients’ attributions of causes can differ from the causes documented in the relevant literature, which is evident in practice, e.g., in the context of patient education and counselling, the generic causes were retained and supplemented by disease-related causes. For example, Rocholl et al. [6] were able to show that the consumption of red wine or pork are among the causes perceived by patients.

Furthermore, eigenvalues > 1 can lead to over-extraction of factors in many situations, especially with larger item sets [32], indicating that this criterion could result in an overestimation of the dimensionality. Future research should verify the number and stability of the factors using more robust methods. Until such confirmatory work is completed, the present factor solution should be interpreted as a provisional, exploratory description of the data.

In addition, no construct validity testing was performed in this study using related instruments (e.g., Hospital Anxiety and Depression Scale or General Self-Efficacy Expectancy Scale [17]). This is recommended for upcoming projects. Convergence and discriminant validity of the assessment instruments have to be verified. Appropriate assessment measures include emotional distress, coping strategies and self-efficacy. Predictive validity should also be examined with respect to therapy adherence and rehabilitation outcomes.

The instrument described, which has been adapted for occupational dermatology, has so far only been used in this form in one institution and has been psychometrically tested for the first time. With 228 respondents, the item-to-participant ratio is approximately 1:6, which is acceptable for a preliminary exploratory study but may limit the stability and replicability of the factor structure, particularly the more complex causal attribution model. Further research should therefore recruit larger samples. A larger sample would improve factor stability and increase the reproducibility of the factor structure in future studies (see [33]). Furthermore, future studies should test the instrument in multiple rehabilitation centres and apply a broader range of validation criteria. This may lead to a broader validation and generalisation of the results. In addition, the “cyclical occurrence” dimension, which could not be replicated in this first test, should be revised in this context. A new sample should be used to confirm the factor structure. In addition, future studies should show the extent to which the instrument can be used to measure change, e.g., to determine the success of interventions in this regard.

In conclusion, it can be stated that this is the first psychometric testing of the IPQ-R for occupational dermatology within a single institution, and that more comprehensive validation in other settings and with larger sample sizes is required before results can be generalised. Considering the internal consistency and the small number of items per scale, the current version is more appropriate for group-level research and the characterisation of cohorts than for decisions involving high risk at the level of individual patients. Further item development—particularly for the weaker subscales—is required before these scales can be used reliably as independent clinical indices.

4.8. Core Messages

The German-language Illness Perception Questionnaire (IPQ-R) modified for occupational dermatology was developed

• On the basis of the existing German-language instrument;
• With an analysis of survey instruments for assessing skin condition;
• Using content-related considerations on expected consequences resulting from a patient survey in health psychology patient seminars during occupational dermatological rehabilitation;
• Based on the results of Rocholl et al. [6].

In the present study, the initial factorial structure of this modified version of the instrument was exploratorily tested for the first time. In its current form, the IPQ-R primarily measures acute or chronic perceptions of hand eczema and does not capture cyclical patterns. Re-development and testing of a dedicated cyclical-timeline dimension constitute a specific target for future research.

Considering the described deviations from the original model, an instrument is available for group-level research and the characterisation of cohorts that was developed as part of this study with a high completion rate and a very good response rate. Both aspects support the assumption that the present IPQ-R version is suitable for use in further research in the field of application.