Validation of a comprehensive Freiburg Life Quality Assessment (FLQA) core questionnaire and development of a threshold system

ARTICLE

Auteur(s) : Matthias AUGUSTIN*, Sabine LANGE*, Kerstin WENNINGER*, Karin SEIDENGLANZ*, Ulrich AMON^†, Ina ZSCHOCKE*

* Department of Dermatology, University Clinics of Freiburg, Hauptstrasse 7, 79104 Freiburg, Germany 
^† Psorisolklinik, Dermatological Treatment Center Hersbruck, Mhlstrasse 31, 91217 Hersbruck, Germany

Article accepted on 15/12/2003

Health related quality of life (HRQOL) has gained substantial interest in most fields of medicine. Several studies have shown that in many skin conditions, HRQOL may be markedly reduced [1-6]. A number of different dermatology-specific HRQOL questionnaires have been used, including the Dermatological Life Quality Index (DLQI) [4] and the Skindex [3] which are the most established and well validated disease-specific HRQOL instruments in the field. However, most of these questionnaires have been developed for use in chronic dermatoses and thus may not be sufficiently applicable to other dermatological conditions such as chronic wounds, chronic vein diseases, herpes, alopecia, or immediate-type allergies. In order to provide an instrument that can be used in different dermatological patient populations but which is also sensitive to specific areas of impact of different skin diseases, we developed a modular HRQOL instrument: the Freiburg Life Quality Assessment (FLQA). Each FLQA questionnaire consists of a core module of generic items as well as a number of items specific to distinct dermatological diseases. That is, specific FLQA versions were developed for different dermatological conditions including the ones specified above. The items of the core module (FLQA-c) are identical for all FLQA versions, permitting direct HRQOL comparisons between different kinds of dermatological affections. 
Another problem of most HRQOL inventories is that they provide absolute data (i.e., average or sum scores) for the different scales. Such scores are difficult to interpret, and results on different questionnaires cannot be compared. For the interpretation of clinical outcome it is therefore advantageous to standardize HRQOL instruments defining a threshold score which indicates a relevantly reduced HRQOL. By calculating the proportion of patients above such a standardized threshold, different patient populations may be compared with regard to the impact of their specific diseases. 
In the present study, the development and validation of the FLQA-c, the generic module that is identical for all FLQA versions, is described. Moreover, the intention of the present study was to define a threshold that indicates the proportion of patients with markedly reduced HRQOL.

Materials and Methods

Development of the Freiburg Life Quality Assessment core module (FLQA-c)
The FLQA-c has been developed in accordance with international principles on the development of HRQOL questionnaires [7, 8] and with the German guidelines for the assessment of HRQOL in dermatology [9, 10]. Item generation was based on a review of existing, well-established generic HRQOL instruments and on open surveys with 56 patients with various dermatologic indications. The questions derived from this item collection were further selected and assigned to the following 6 a-priori scales: Physical complaints, everyday life, social life, emotional status, stress due to treatment, and satisfaction with different areas of life. Except for the items on stress due to treatment, all questions were phrased in a way applicable to patients as well as to healthy persons. The resulting FLQA-c comprises 28 items (see appendix). The scoring procedure includes recoding several items of the scale emotional status. After recoding, higher values represent a lower HRQOL on all items. Answers within a scale are averaged to generate a scale score, resulting in a scale score range from 1 to 5. The measure does not provide a total score.
For the development of specific versions of the FLQA, particular item collections in the targeted patient samples were made. The resulting series of questionnaires including the core items as well as disease-specific items were tested in separate validation studies reported elsewhere [1, 2]. Versions for ten different dermatological indications are under development.

Patients and study procedure

For the validation study, 655 inpatients (n = 401 with psoriasis, n = 254 with atopic dermatitis) from the Psorisol Dermatological Treatment Center Hersbruck, Germany, were recruited. 98.8% (n = 647) of the patients completed and returned the questionnaires. A sub-sample of 378 patients filled out the FLQA-c a second time after four weeks of inpatient treatment. These data were used to assess the instrument’s sensitivity to change. Clinical scores were taken from the patients’ clinic charts. For psoriasis patients, the Psoriasis Area and Severity Index (PASI) [11] was used as clinical score, for atopic dermatitis patients, the Scorad [12] was used. In addition, a sample of 240 control subjects from the general population participated in the study. Most of the control subjects were recruited in several German clerical community centres. The rest of the control subjects were recruited among medical students at the University of Freiburg. In addition to the FLQA-c, all participants completed two further HRQOL instruments described below. Socio-demographic and clinical data are displayed in Table I. Due to the large sample size, significant differences between the three groups were found for age, duration of school education, and sex distribution. On average, the atopic dermatitis patients were significantly younger than the controls, who in turn were significantly younger than the psoriasis patients (F_(2,865) = 52.5, p <.001). On average, the psoriasis patients’ duration of education was significantly shorter than that of the atopic dermatitis patients, and it was longest for the controls (F_(2,865) = 59.9, p <.001). These differences are not of theoretical concern, however, because age and duration of education were only minimally correlated with the FLQA-c scales (range of r for the different FLQA-c scales: –.17 to +.08). Consequently, analyses of covariance including age and duration of education as covariates yielded exactly the same conclusions as the analyses of variance reported below. Significant differences were also found regarding the sex distribution (Chi²₍₂₎ = 51.1, p <.001): the percentage of male participants was lowest in the group of atopic dermatitis patients, and highest in the group of psoriasis patients. These differences do not compromise our conclusions either: correlations of sex with the FLQA-c scales were low (between 01 and 21), and analyses of variance including sex as a factor yielded a significant interaction with group status only for the “emotional status” scale (the difference between the two patient groups was slightly more pronounced in males). In summary, these results indicate that the small, but significant differences found for age, duration of school education, and sex distribution do not compromise the conclusions reported below. Table I also shows that there was a wide range of severity in both patient groups (see range of clinical scores in bottom row). However, we found only low and moderate correlations between these scores and the FLQA-c scales (between r = 10 and r = 22 in the psoriasis patients, between r = 14 and r = 34 in the atopic dermatitis patients). This result suggests that health related quality of life may be fairly independent of objective severity of skin diseases.

Table I. Socio-demographic and clinical data of the patient and control samples

Validation procedures

Reliability: Internal consistency

Internal consistency coefficients of the scales were determined based on the FLQA-c questionnaires of the whole study population.

Discriminant validity. FLQA-c scale score differences between psoriasis patients, atopic dermatitis patients, and the control sample were assessed as a measure of discriminant validity.

Convergent (construct) validity

Construct validity was determined by correlating the results of the FLQA-c with corresponding scales of the following two questionnaires: 1) The Dermatological Life Quality Index (DLQI) [4] as a well-established disease-specific HRQOL measure (an authorized German translation of this questionnaire was used) and 2) the Questionnaire on Everyday Living (ALLTAG) [13] as a generic HRQOL questionnaire frequently used in the German language area.

Sensitivity to change

Sensitivity to change was evaluated by a pre-post-comparison of the patient groups on the FLQA-c scales before and after four weeks of inpatient treatment. The therapy program integrated standard dermatological treatment, patient education, and behavior-oriented psychotherapy.

Thresholds indicating patients with markedly reduced HRQOL

Based on the FLQA-c data distributions of the control sample, a threshold value indicating a markedly reduced HRQOL was defined. The cut-off was set at one standard deviation above the mean, i.e. values greater than one standard deviation above the mean were defined as indicating a significant reduction in HRQOL. The thresholds determined in this study are shown in the final column of Table III. In support of this procedure, the authors argue as follows:
a) In a distribution not differing substantially from a normal distribution, a cut-off point at one standard deviation above the mean detects about 17% of the sample. As earlier epidemiological data [14] have revealed, about 20% of persons in an unselected sample of the general population showed marked psychological distress. Thus, the proposed cut-off value detecting about 17% of our control sample as having a reduced HRQOL is roughly in accordance with epidemiological data on the general psychological morbidity.
b) A number of previous studies [15-17] have defined a threshold for significant disease-related stress in patients with chronic dermatoses and skin tumors using the same methodological approach. The statistically based thresholds were highly concordant with clinical ratings and are thus considered to be a valid method.
c) In general, using distribution characteristics to define a threshold is a common approach in interpreting clinical outcome data. e.g., the concept of clinical significance [18] which aims at defining and standardizing a clinically meaningful size of observed treatment effects and is also based on a cut-off score calculated from the mean and standard deviation of the outcome parameter.

Statistical analyses

All computations were calculated with the Statistical Package for the Social Sciences (SPSS). Comparative tests between the three groups were performed using analyses of variance for independent samples, followed by alpha-controlled post-hoc tests (Tukey’s HSD). T-tests for paired samples were employed for pre-post treatment comparisons. As a measure of concordance, Pearson’s correlation coefficients were computed. For reliability analyses, Cronbach’s alpha coefficients were used, yielding measures if internal consistency. All statistical comparisons were two-sided, employing p <.05 as the level of significance. Where multiplicity was present, the Bonferroni correction (i.e. dividing the alpha level of.05 by the number of tests) was implemented to avoid inflation of the alpha error.

Results

Distribution characteristics and internal consistencies

Except for the scale “treatment”, Cronbach’s alpha coefficients were satisfactory for all scales with values of 0.75 or above. There were no expressed floor or ceiling effects with the exception of slight bottom effects in the scales “social life” and “everyday life”. (Table II)

Discriminant validity

The FLQA scales showed very good discrimination between the three groups of psoriasis patients, atopic dermatitis patients, and control subjects (Table III). Except for the scale “treatment”, the atopic dermatitis patients reported significantly lower HRQOL than patients with psoriasis on all scales. Compared to the control sample, patients of both diagnoses reported significantly lower HRQOL on all scales.

Convergent (construct) validity

Corresponding scales of the FLQA-c and the DLQI [4] and the ALLTAG [13] showed significant correlations between 0.33 and 0.65 (Table IV). While there was only a low correlation between the ALLTAG scale “social life” and the FLQA-c scale “social life”, the FLQA-c scale was moderately correlated with the DLQI scales “spare time/sports” and “relationship”. In summary, the convergent validity of the FLQA-c scales “physical complaints”, “everyday life”, “social life”, “emotional status”, and “satisfaction” was satisfactory, since correlations with comparable scales of the ALLTAG and the DLQI were moderate. The correlation between the FLQA-c and DLQI scales “therapy” was also moderate, and therefore still satisfying.

Sensitivity to change

Satisfactory sensitivity to change was found for all scales (Table V) in both patient groups. After four weeks of integrated dermatological and psychosomatic treatment, both psoriasis patients and atopic dermatitis patients reported increased HRQOL in all domains.

Patients with HRQOL impairments above threshold

No severe deviations from a normal distribution were detected for the FLQA-c scale scores in the control sample (skewness and kurtosis values were below 1.0 for the majority of the scales). The score distributions for “everyday life” and “social life” were slightly L-shaped but still had acceptable skewness values – 2.0 and 1.7, respectively – and kurtosis values – 4.2 and 2.7, respectively. On the different scales, 13% to 15% of the control subjects were detected by the threshold of one standard deviation above the mean as having HRQOL reductions (Fig. 1). As expected, proportions of persons with reduced HRQOL were markedly higher in the patient samples. These results indicate that for each scale, about half of the patients (approx. 54%) suffer from quality of life reductions which are so severe that only 13-15% of the control subjects experience them. The highest percentages (about 60-70% for the different scales) were found for patients with atopic dermatitis.

Discussion

The aim of this study was to validate the FLQA-c questionnaire, the core part of a modular HRQOL instrument that permits direct HRQOL comparisons between different kinds of dermatological diseases. The development of the questionnaire was performed according to international principles of test development and validation [7, 8] and was in line with the German guidelines [9, 10] for the assessment of HRQOL in dermatology.
The results support the FLQA-c as a reliable and valid instrument for the assessment of HRQOL. The data suggest that it is applicable to patients with chronic dermatoses as well as to the general population, and that it may be used in cross-sectional clinical studies as well as in longitudinal trials evaluating treatment induced change. As a major advantage over previous HRQOL instruments available for dermatological patient samples, the FLQA-c may be combined with one of ten modules specific to certain skin diseases [1, 2], which enhances the discriminative power and sensitivity of the questionnaire.
Especially in pharmaco-economic studies, the option of directly comparing different dermatological conditions with one identical HRQOL instrument may be beneficial.
Our proposed threshold system, which is based on distribution characteristics to define a cut-off, identified a proportion of 13% to 15% of the control subjects as having HRQOL reductions. This is a more conservative estimate of the proportion of the general population experiencing a relevant level of distress than has been reported in earlier epidemiological studies [14]. However, despite this conservatively estimated cut-off point, the proportion of patients identified as experiencing a decreased HRQOL was strikingly high. These clinical results of expressedly reduced HRQOL in our patient populations are in accordance with previous reports by Finlay [4] and Chren [3]. We would like to emphasize that the measurement of HRQOL is an indirect approach since HRQOL as a “quality” cannot be quantified. Hence, it is difficult to interpret absolute data on HRQOL scales. However, the detected high proportions of patients with marked impairments of HRQOL suggest that patients with atopic dermatitis and psoriasis should routinely be checked for specific HRQOL reductions. In the case of severe strain, psychosocial support should be considered and made available to the patient.
In summary, there are three potential objectives of a standardized threshold detecting patients with reduced HRQOL: First, it provides a point of reference for the interpretation of absolute scores of individual patients. Second, results on different HRQOL questionnaires may be set in relation to each other. For instance, one instrument may detect a higher proportion of a clinical sample as having a reduced HRQOL than a second questionnaire, which indicates that the first measure is more sensitive to the specific impact of the disease. Finally, different patient populations may be compared with regard to the percentage of patients significantly affected by their disease. Knowing the proportion of patients with significantly impaired HRQOL in different skin disease populations permits a more precise allocation of psychosocial as well as clinical resources. Thus, the detection of “risk groups” of severely affected patients may help saving costs while maintaining or improving the quality of care.
A number of limitations of this study have to be considered. First, our clinical sample included only patients with chronic dermatoses. Further research is needed to explore whether the psychometric properties of the FLQA-c are as satisfying when assessed in other dermatological patient samples. Also, the test-retest reliability needs to be assessed in future studies since no retest data were collected from patients whose health status remained stable. Finally, since a selection effect can never be excluded for volunteer control samples recruited from selected groups of the population, it remains open to what degree our control sample was representative of the general population. Ideally, the threshold identifying persons with a significantly reduced HRQOL as well as norm values for comparison should be recalculated from a large, representative normal population sample. n

References

1. Augustin M, Dieterle W, Zschocke I, Brill C, Trefzer D, Peschen M, Vanscheidt W. Development and validation of a disease specific questionnaire on the quality of life of patients with chronic venous insufficiency. VASA 1997; 26: 291-301.

2. Augustin M, Zschocke I, Seidenglanz K, Lange S, Schiffler A, Amon U. Validation and clinical results of the FLQA-d, a quality of life questionnaire for patients with chronic skin diseases. Dermatol Psychosom 2000; 1: 12-7.

3. Chren MM, Lasek RJ, Quinn LM, Mostow EN, Zyzanski SJ. Skindex, a quality of life measure for patients with skin disease: reliability, validity and responsiveness. J Invest Dermatol 1996; 107: 707-13.

4. Finlay AY, Kahn GK. Dermatology Life Quality Index (DLQI): a simple practical measure for routine clinical use. Clin Exp Dermatol 1994; 9: 210-6.

5. Lundberg L, Johannesson M, Silverdahl M, Hermansson C, Lindberg M. Health-related quality of life in patients with psoriasis and atopic dermatitis measured with SF-36, DLQI and a subjective measure of disease activity. Acta Derm Venereol 2000;80:430-4.

6. Zachariae R, Zachariae H, Blomqvist K, Davindsson S, Molin L, Mrk C, Sigurgeirsson B. Quality of life in 6497 Nordic patients with psoriasis. Br J Dermatol 2002; 146: 1006-16.

7. Finlay AY. Quality of life measurement in dermatology: a practical guide. Br J Dermatol 1997; 136: 305-14.

8. Sackett DL, Chambers LW, MacPherson AS, Goldsmith CH, McAuley RG. The development and application of indices of health: general methods and summary of results. Am J Public Health 1977; 67: 423-8.

9. Augustin M, Amon U, Bullinger M, Gieler U: Recommendations for the assessment of quality of life in dermatology. Dermatol Psychosom 2000; 1: 84-7.

10. Augustin M. Erfassung von Lebensqualität in dermatologischen Studien. Leitlinie der Subkommission “Pharmako-Ökonomie und Lebensqualität” (Recording of quality of life in dermatological studies. Guidelines of the subcommission “Pharmacoeconomics and quality of life”). Hautarzt 2001; 52: 697-700.

11. Frederiksson T, Pettersson U. Severe Psoriasis – Oral Therapy with a new Retinoid. Dermatologica 1978; 157: 238-44.

12. European Task Force on Atopic Dermatitis. Severity scoring of atopic dermatitis: The SCORAD Index. Dermatol 1993; 186: 23-31.

13. Bullinger M, Kirchberger I, von Steinbchel N. Der Fragebogen Alltagsleben – ein Verfahren zur Erfassung der gesundheitsbezogenen Lebensqualität. (The Questionnaire on Everyday Living – an instrument for the assessment of health-related quality of life). Z Med Psychol 1993; 3: 121-31. 

14. Schepank H. Epidemiologie psychogener Erkrankungen – Ein Beitrag zur Grundlagenforschung aus einer Feldstudie (Epidemiology of psychogenic diseases – A research contribution from a field study). Z Psychosom Med 1982; 28: 104-25.

15. Fritzsche K, Ott J, Zschocke I, Scheib P, Burger T, Augustin M. Psychosomatic liaison service in dermatology. Need for psychotherapeutic interventions and their realization. Dermatology 2001; 203: 27-31.

16. Strittmatter G, Mawick R, Tilkorn M. Psychosozialer Betreuungsbedarf bei Gesichts- und Hautturmorpatienten. (The need for psychosocial intervention in facial and skin tumor patients). Psychother Psychosom med Psychol 1998; 48: 349-57.

17. Strittmatter G, Tilkorn M, Mawick R. Screening of patients for psychosocial risk by the doctor - a contribution to assurance of quality of treatment integrated intervention. Psycho-Onkology 1998; 7(Suppl.): 142.

18. Jacobson NS, Roberts LJ, Berns SB, McGlinchey JB. Methods for defining and determining the clinical significance of treatment effects: Description, application and alternatives. J Consult Clin Psychol 1999; 67: 300-7.

Appendix 

Freiburg Life Quality Assessment – Core version (FLQA-c)*

This questionnaire serves for the evaluation of your health related quality of life. It refers to various areas of life.

Please respond to the questions carefully, but spontaneously. All responses will be confidential and entered into the database anonymously.

1. Physical complaints

The following questions are related to your physical well-being.
Please mark the appropriate box.
How often did you notice in the previous week...

2. Daily life

The following questions address your daily life.
For each statement, please mark the appropriate
box best describing your situation during the last week...

3. Social life

The following questions address your relationships with others and the impact your disease (if you were ill) may have on them.
Please mark the appropriate box.
How often did you notice in the previous week...