The aim of this study was to develop a new specific instrument that would be able to measure satisfaction with treatments for hand eczemas, especially CHE, to test its psychometric properties, and to provide data regarding its validity.
The findings of this study show that the DermaSat Questionnaire is a valid, reliable, and feasible instrument for use in routine medical practice, both as a unidimensional instrument for comparing patients (using the total score), and when the clinician wishes to explore patient satisfaction with the various facets of treatment (for which the subscales of the instrument have also proven valid and reliable).
CHE is a disease that is difficult to treat, which can severely impact quality of life. As demonstrated in the literature, the impact can be as severe as in asthma or psoriasis . Currently, few effective therapeutic options are available to help patients. Furthermore, the impact of CHE on patients' lives is underestimated. Clearly, CHE is not life-threatening but, since it is centered on the hands, it may affect relationships and social integration, and even cause rejection.
Several treatment satisfaction questionnaires have been developed to date and may be used in patients with CHE. Due to the very specific circumstances of CHE, none of the existing instruments was deemed specific enough to give clinicians a reliable tool for evaluating treatment satisfaction. It was felt that there was a clear need to develop a specific CHE treatment satisfaction questionnaire. The specificity of this questionnaire will allow the clinician to better understand the patient's personal experience with current treatments, based on an objective tool. Depending on the outcomes, the clinician will be able to change and adapt the treatment algorithm.
The results obtained reveal that the DermaSat has very good metric properties. From the standpoint of feasibility, the response rate is highly satisfactory (nearly all patients answered all questions) and the time required to administer it was brief (under eight minutes on average), which makes it highly feasible for use at any level of health care, particularly at the primary level, where the time available for care is usually short.
With regard to the reliability of the questionnaire, both internal consistency and two halves reliability methods attain higher values than the accepted standard minima  with respect to the total score and also to the individual subscales. And the analysis of the isolated items indicates that all of them have good discriminant capabilities.
The various aspects of the validity of the questionnaire that were analyzed produced satisfactory results. The content validity (originally established by a panel of experts) was confirmed by the resulting factor structure. Furthermore, the study of the responses through exploratory and confirmatory factor analysis corroborated the initially proposed theoretical structure and specifically corroborated the presence of six subscales or dimensions: treatment effectiveness, convenience of use, impact on daily activities, medical care, undesired side-effects, and general satisfaction. The relationship detected between the various dimensions suggests that the scores of the different subscales can be combined into a meaningful total score. Whereas we observed a close relationship between the dimensions of effectiveness, impact on daily living, and general satisfaction in the confirmatory analysis, the correlation of the various dimensions with other assessments by patients and the clinician responsible for the treatment suggests that they could actually be considered as different dimensions. It would however be advisable to validate the result in new and larger patient samples.
Multidimensional models similar to the one obtained here have been successfully applied to generic medicine treatments (SATMED-Q)  and treatments for asymptomatic diseases such as glaucoma (GLAUSAT) , and the dimensions obtained have been shown to be consistent with other models such as the TSQM [18, 24]. As a benchmark, we found that the correlation of overall SATMED-Q and TSQM scores was high and positive (r = 0.74), and convergent dimensions of both questionnaires correlated between 0.58 and 0.68.
Comparing the structure obtained for dermatologic treatments with the structure obtained in generic treatments and in glaucoma treatments, several interesting differences are worth mentioning. The number of dimensions and their composition are similar to those obtained for generic treatments, while in asymptomatic treatments as glaucoma an additional dimension covering expectations and beliefs about treatment was needed. Correlations between dimensions (their observed proximity) vary in the three instruments, and we will discuss some of the differences.
In dermatologic treatments Treatment Effectiveness is more related to the other dimensions in the questionnaire. In the DERMASAT, Treatment Effectiveness highly correlates with General Satisfaction (0.92) while the correlation is much lower in generic (0.76) and glaucoma (0.67) treatments. Treatment Effectiveness also highly correlates with Impact on Daily Life (0.85) as compared to generic (0.75) and glaucoma (0.09) treatments; it has a higher correlation with Medical Care (0.53) than generic (0.38) and glaucoma (0.33) treatments; a similar correlation with Convenience of Use (0.43) to generic treatments (0.41), but higher than glaucoma (-0.19); and a similar correlation with Undesired Effects (-0.18) to generic (-0.12) and glaucoma (-.16).
Convenience of Use exhibits a somewhat lower correlation with Impact on Daily Life (0.39) than in generic treatments (0.42) but much lower than in glaucoma (.55); correlates higher with Medical Care (0.42) than in generic (.20) and glaucoma (0.00) treatments; and correlates similarly with General Satisfaction (0.43) to generic treatments (0.45) and higher than in glaucoma ( 0.05).
Undesired effects shows a higher correlation with Impact on Daily Life (-0.24) than in generic ( 0.11) and glaucoma (-0.17); lower correlation with General Satisfaction (-0.19) to generic (-0.32) and glaucoma (-0.25); and similar with Medical Care (-0.10) as in generic (-0.12), but lower than in glaucoma (-0.26).
Medical Care correlates highly with Impact on Daily Life (0.52) as compared to generic (0.32) and glaucoma (0.30) treatments; and also with General Satisfaction (0.59), as compared to generic (0.35) and glaucoma (0.28).
Summarizing, the structure of patient treatment satisfaction exhibits particular characteristics that need to be considered. Treatment effectiveness is much more related to other aspects of patient satisfaction than in generic and asymptomatic treatments, but it is not confused with undesired effects. Additionally, satisfaction with medical care is also more related to other aspects of satisfaction and its relation with convenience of use and impact on daily life is over-weighted. Undesired effects are clearly distinguished but related to impact on daily life.
From previous results, we expected moderate correlations between satisfaction dimensions and patient reports of compliance, except with undesired effects and medical care. Instead, convenience of use is the only dimension found to have a correlation close to significance with this additional treatment outcome. Hence, usability seems to be the only aspect of treatment satisfaction that could help to forecast treatment compliance. More consistent with theory is the observed correlation between satisfaction with treatment effectiveness and compliance assessed by the clinician. In fact, the correlation between patient and clinician assessment of compliance is only moderate (-0.30), but higher than values obtained for glaucoma treatment (-0.09). These results might reflect the fact that accurate and valid measures of self-reported compliance have yet to be developed, and we will have to wait until such instruments are available before being able to study in detail the relationship between compliance and treatment satisfaction.
On the other hand, effectiveness and tolerability (both assessed by the clinician and the patient) do correlate with satisfaction dimensions and with the overall satisfaction score. Overall satisfaction correlates with clinician and patient effectiveness at a level similar to previous generic questionnaires (0.41 and 0.61, respectively) and higher than in glaucoma (0.27 and 0.19, respectively). Correlations of treatment satisfaction with clinician and patient tolerability assessments are however lower than those found for generic instruments (0.26 and 0.39) and much lower than those found in glaucoma (0.44 and 0.35). Treatment effectiveness seems to be the most influential aspect related to patient satisfaction, while tolerability is a much milder explicative factor.
Differences found between gender groups were not foreseen in advance due to the exploratory nature of the comparison. Nevertheless, they are consistent with other findings from the focus group discussions. Women are more used to regular care of their hands, and don't find it inconvenient to use creams in their daily living, while men (at least in our culture) are not used to it or even show prejudices against using creams.
We expected that more severe disorders would reflect a lower level of satisfaction but results are not conclusive and additional research would be needed. Although the general pattern of means gives some support to this idea (see Figure 2), differences do not reach significance. This could be due to not having enough statistical power owing to small sample size. In fact, 5 out of 9 groups contained fewer than 20 patients, while mean differences are larger than those found between gender groups. But it also could be justified by the fact that there is a very wide range of different treatments in our sample (more than 36 different treatment profiles) and an even wider range of effectiveness experienced by patients.
Regarding concurrent validity, only some of the DermaSat scores correlate with the quality of life and state of health assessments by the patients themselves, suggesting that treatment satisfaction is actually a different type of divergent construct. No relationship is found between the treatment satisfaction assessments and the compliance assessments, something that was not expected in principle, and which requires further research with other groups of patients. Regarding construct validity, we found convergent results when we compared the DermaSat scores with the effectiveness and tolerability assessments (by the clinician and the patient): the greater the perceived effectiveness, the higher the satisfaction scores; the greater the perceived tolerability, the higher the medication-related satisfaction.
As with the SATMED-Q, the DermaSat presents two dimensions of special interest for the patient, one assessing the impact of the medication on activities of daily living and the other assessing patient satisfaction with medical care. These two aspects are highly valued by patients and can serve as an aid to clinicians in making decisions about treatment.
One limitation of this study is that it was cross-sectional and was not capable of examining casual influences of low treatment satisfaction on clinically relevant outcomes. In addition, the responsiveness of this questionnaire to change was not examined in this cross-sectional study. Prospective studies are being planned to address this issue. Another limitation of this study is the fact that no generic instrument measuring satisfaction with treatment was used concurrently. This task was not performed in order to avoid overloading the patients who came to a regular appointment and who were not compensated for their participation in the study. We believe it would be advisable to gather additional data in this regard, a matter that should be easy to tackle now that we have a short questionnaire.
A final limitation of the DermaSat is that the primary validation samples were enrolled in Spain; therefore, the validation of this instrument in international settings is unknown and must be tested.
Our hands are our main 'working tool' and, in some respect, our 'business card' when meeting other people.
The results of our study suggest that treatment satisfaction is very low in this kind of condition, independent on the severity of lesions. All hand lesions, especially if the finger-tips are involved, can severely handicap a patient's daily living when there is no or very low treatment response. This is true for both severe and moderate CHE. This particular combination translates into very low treatment satisfaction with current treatments as found in our study.
On top of this, from a clinician's point of view, correct classification of CHE severity is quite subjective.
It is often the case that mild to moderate hand lesions become chronic and therefore result in severe CHE, even though the lesions per se might not 'look so bad'.