The study sample included 983 women from 101 dermatology clinics in the different Spanish autonomous communities. The women consulted for female androgenetic alopecia and participated voluntarily after receiving appropriate information. They started treatment with dietary supplements containing 5-α-reductase enzyme inhibitor and had not received treatment with supplements in the previous 3 months.

Quality of life was evaluated using HSS-29 (Appendix B1. Additional Material) and SF-12 (Appendix B2. Additional Material). The HSS-29 questionnaire is a self-administered tool that makes it possible to evaluate HRQOL in patients with alopecia. It was based on the Skindex-29 scale and is composed of 29 items distributed over 3dimensions or domains: symptoms (7items), function (12items), and emotions (10items). The questionnaire takes a mean of 5minutes to complete, and each item is answered based on a Likert-type scale ranging from 0 (never) to 5 (all the time). The dimensions are scored by transforming the sum of all the responses to a linear scale ranging from 0 (no impact) to 100 (maximum impact).

The SF-12 questionnaire contains 12items covering 8 HRQOL domains: physical dimension, mental dimension, physical functioning, role-physical, bodily pain, general health, role-emotional, social functioning, and mental health. Reference guidelines are available for the Spanish version of SF-12 (version 2).4

Furthermore, the patient and the clinician were asked their opinion on the volume, density, and texture of the hair.

Patients attended 2 visits, one at baseline and the other after 6months of treatment. At the baseline visit, patients initiated treatment with 5-α-reductase enzyme inhibitor dietary supplements, which they continued to take for 6months. They completed both questionnaires at each visit, and the investigators evaluated the severity of alopecia using the Sinclair scale. Both the patient and the clinician evaluated the volume, density, and texture of the hair as better, worse, or no change.

The project for cross-cultural adaptation to Spanish began in 2016, according to the protocol approved by the Ethics Committee of Hospital 12 de Octubre on December 20 (No. 16/392). The second phase was carried out between January and October 2018.

First, a descriptive analysis of the data was performed using a frequency table for nominal variables and measures of central tendency and dispersion for continuous variables. Data for continuous variables were presented with their 95%CI.

Comparisons between groups and quantitative variables were made using the t test (2groups) or 1-way analysis of variance (≥3 groups). If either of these tests could not be applied, the Mann-Whitney test (2groups) or the Kruskal-Wallis test (≥3 groups) was used. Where necessary, tests for paired data were applied. Continuous variables were compared using the Fisher correlation coefficient in the case of absolutely continuous variables or the Spearman test in the case of those that were not. All calculations were made using SAS Version 9.4.

Table 1 shows the characteristics of the patients included in the study. The study population comprised 983 women with a mean age of 50 (16.3) years and mainly (83.6%) grade 1–3 alopecia according to the Sinclair scale. Table 2 shows the total score and the score by dimensions on the HSS-29 scale. The grade of alopecia is shown in Fig. 1. Emotion was the most affected domain, with a mean score of 37.7 (95%CI, 36.3–39.2); the score was lower in the symptom and function domains, with a mean of 27.2 (95% CI, 26.0–28.4) and 19.3 (95%CI, 17.9–20.6), respectively. We found that involvement was more pronounced as alopecia became more severe, both in the overall score and in the various dimensions. No significant differences were detected for age group with respect to the score on the scale as a whole or in its individual dimensions. The baseline SF-12 score is also shown in Table 2. The domain with the greatest impact on quality of life was role-emotional, with a mean score of 20.3 (95%CI, 20.0–20.6). As in the HSS-29, involvement was more pronounced as alopecia became worse in all the scores recorded for the SF-12.

Figure 1.

Hair-Specific Skindex-29 and severity of alopecia.

Total score and score by dimension according to the severity of alopecia (Sinclair scale) at the baseline visit. Involvement in the function, emotion, and symptom dimensions of the Hair-Specific Skindex-29 scale and overall involvement increase with the severity of alopecia.

t test/ANOVA between the score on the function dimension and grade of severity: P<.001. t test/ANOVA between the emotion dimension and grade of severity: P<.001. t test/ANOVA between the symptom dimension and grade of severity: P<.001. t test/ANOVA between the overall score and grade of severity: P<.001. ANOVA indicates analysis of variance.

(0.25MB).

The correlation between both scales at baseline is presented in Table 3. All of the correlations studied using Pearson coefficients were negative and statistically significant (P< .001). A higher score in the HSS-29 and all its subdimensions (function, emotion, and symptoms) was associated with a lower score, i.e., reduced HRQOL, in all of the dimensions of the SF-12 scale. Similarly, the mean scores on the SF-12 decreased as severity in the HSS-29 increased in its classified form. When the study population was stratified by age group and severity of alopecia, we observed the same negative correlation in all the groups, except in patients with grade 5 alopecia. No analysis was performed for this subgroup, since it only included 3women.

Of the 983 patients included in the initial project, 967 (98.3%) were evaluated at 6months. Compared with the baseline score, the global score on the HSS-29 scale at 6months decreased overall and in the scores for the individual dimensions, in which the differences were statistically significant (Table 4). The mean score on the overall scale fell from 27.5(18.7) at baseline to 19.3(15.7) at the follow-up visit. Significant changes were also observed for the function, emotion, and symptom dimensions, from 19.3(2.0) to 13.1(15.8), from 37.7(23.0) to 25.8(20.5), and from 27.2(18.8) to 20.7(16.2), respectively. Fig. 2 shows the differences between the 2 visits according to the grade of alopecia. All of the differences evaluated were statistically significant with respect to the grade of alopecia.

Figure 2.

Hair Specific Skindex-29.: Differences in score between the first and second visit for each grade of alopecia. The presence of negative differences indicates a reduced impact on quality of life at the second visit.

Difference at 6months from baseline in the function dimension: t test/ANOVA P<.001. Difference at 6months from baseline in the emotion dimension: t test/ANOVA P<.01478. Difference at 6months from baseline in the symptom dimension: t test/ANOVA P<.0009. Difference at 6months from baseline in the Hair Specific Skindex-29 global: t test/ANOVA P<.0001.

(0.25MB).

Depending on the severity of alopecia, differences were observed in the score on the HSS-29 scale between the first and second visits, with an improvement in involvement of quality of life (i.e., scores <20 in the HSS-29 scale in all its domains and for all grades of alopecia). Fig. 2 shows the differences between the first and second evaluation.

A statistically significant increase in the SF-12 scores was observed in the dimensions role-physical, bodily pain, general health, vitality, role-emotional, social functioning, and mental health at 6months, whereas the change in the physical dimension and physical functioning was not significant (Table 4).

In order to evaluate sensitivity to change, we calculated the differences between the scores at 6months and the baseline scores for all the dimensions on the HSS-29 and SF-12 scales and assessed the evaluations of the appearance of the patients’ hair. We then determined whether these differences were correlated or not, that is, whether one changed with the other. The analysis was performed for the whole study population and by age group and grade of alopecia.

Table 5 shows the correlation for the differences between the scales, that is, whether one changed with the other. We observed that the differences between baseline and follow-up both for the overall HSS-29 index and for each of its dimensions were significantly correlated with the differences found for the dimensions of the SF-12. The Pearson correlation coefficients ranged between −0.1 and −0.4, and statistical significance in all cases was P<.001. The scatterplots for the correlation of differences are shown in Appendix BAnnex 3 (Additional material).

Lastly, Table 6 shows the differences between the appearance of the patients’ hair and the HSS-29 score. All of the correlations examined were statistically significant, except for texture as assessed by the clinician compared with progress in the emotion dimension. As the indices for the evaluation of the hair worsened, the HSS-29 score decreased. Both the physician’s and the patient’s evaluation revealed an improvement in the appearance of the hair in the categories volume, density, and texture. For a score of 1–3 (1 normal, 3 very diminished), the means perceived by the patients decreased from 2.3 (0.5) to 1.8 (0.6), from 2.3(0.5) to 1.8(0.6), and from 2.0(0.7) to 1.5(0.6) in volume, density, and texture, respectively. The means perceived by the physician decreased from 2.0(0.5) to 1.5(0.6), from 2.1(0.5) to 1.7(0.5), and from 1.8(0.6) to 1.3(0.5). All of the results were significant. In approximate terms, the higher the grade of alopecia, the greater the improvement, although data for grade5 alopecia were not taken into consideration because there were only 3 patients in this group. No statistically significant differences in patient progress were detected by age group.

Evaluation of the psychometric properties of an instrument is essential when determining quality of the measurements it generates. When evaluating the accuracy of an instrument, the key characteristics of a measurement are reliability, validity, sensitivity, and feasibility. Reliability refers to the ability to measure a variable in a constant manner; validity refers to whether or not the instrument measures what it is intended to measure. Not all reliable instruments are valid, since an instrument may be reliable because it measures the variables in a constant manner but invalid if it does not appropriately measure the phenomenon it is intended to measure.5 The validity of an instrument also depends on its sensitivity and feasibility. Sensitivity shows the ability to detect changes in items or subjects evaluated after an intervention.6 It is associated with differences in the scores of a subject who has improved or worsened and scores that remained unchanged.7 Lastly, feasibility evaluates whether the questionnaire can actually be used in the field it is intended to be used in.

Both generic HRQOL questionnaires and dermatology-specific questionnaires have been used for the study of quality of life in many skin diseases.8 In patients with alopecia, Williamson et al.9 used an adapted version of the Dermatology Life Quality Index, although this is not specific for measuring the effect of alopecia in women, which is probably more intense than in men. In the case of women with alopecia, Schmidt et al.10 applied the Hairdex scale (not available in Spanish), and, in Finland, Hirsso et al.11 used the generic RAND36 scale. The Symptom Checklist-90-R, which is specific to psychopathology, has also been used as a complementary approach.12

HSS-29 is the only specific instrument available in Spanish for evaluation of HRQOL in women with alopecia2. In the present project, we completed the validation of HSS-29. Our data show that it is sensitive to change, that is, the score changes as the grade of female alopecia decreases. Similarly, the values obtained in the HSS-29 are significantly associated with the aspects evaluated using a generic HRQOL scale.

The correlations between the scales were statistically significant at both the baseline visit and at the follow-up visit. Quality of life in all the dimensions of SF-12 worsened as the score on the HSS-29 and its subdimensions (function, emotions, symptoms) increased. Thus, the correlations are negative, because in the case of HSS-29, the score for each dimension is obtained by transforming the sum of all the responses on a linear scale of 0 (no impact on quality of life) to 100 (maximum impact), whereas on the SF-12, a higher score indicates better quality of life. Involvement in the emotion domain of the HSS-29, with a mean score of 37.7, was higher than in the symptom and function domains. This is consistent with the SF-12, where role-emotional was the dimension most affected by alopecia. The instrument seems to be very sensitive, since, compared with baseline, the HSS-29 score at 6months decreased both overall and in the individual dimensions, for which the differences were statistically significant. This aspect is relevant, both in the validation of the test itself, since sensitivity is an important psychometric property, and in the evaluation of the efficacy of therapy.

Women with female androgenetic alopecia have a more negative perception of their health and psychosocial situation than women who do not experience hair abnormalities.10,13,14 The condition may even be associated with depressive symptoms.15 Van der Donk et al.16 reported that 50% of women with alopecia had social problems and that 80% felt that their daily life was affected to some extent. In a study performed in Finnish women, Hirsso et al.11 suggested that alopecia may even be associated with physical symptoms. We found that the effect of female androgenetic alopecia on HRQOL was considerable, particularly in the emotion domain: the more severe the alopecia, the more the patient was affected, both overall and in the individual dimensions. This impact is reflected in data obtained using the HSS-29 and the SF-12, although the former was more sensitive to change. The SF-12 focuses on physical limitations and emotional status independently of the cause, and it may be for this reason that the HSS-29, which focuses more on the impact of alopecia, is more sensitive to changes in the condition of the hair.

The validation project is particularly interesting because of its sample size (983 women), which is larger than that of other studies on quality of life in dermatology. In addition, the number of losses was relatively small (12.1%). Similarly, questionnaires on quality of life translate concepts to numerical scales, although this measurement could be completed with a qualitative evaluation. A growing number of researchers now use such an approach.17