The Influence of Atopic Dermatitis on Health-Related Quality of Life in Bangladesh

Atopic dermatitis (AD) is the foremost non-fatal skin-related disease that affects all age groups. Despite the growing prevalence of AD in low- and middle-income countries, its physiological consequences remain overlooked in countries like Bangladesh. Therefore, we aim to assess and characterize the influence of AD on the health-related quality of life (HRQoL) in Bangladeshi patients. A cross-sectional study comprising 184 eligible adults (83 men and 101 women; mean age, 33.46 ± 15.44 years) was conducted at the dermatology outpatient department of Shaheed Suhrawardy Medical College Hospital (a tertiary hospital in Dhaka, Bangladesh). AD was determined using the UK Working Party criteria. A structured questionnaire, Eczema Area and Severity Index (EASI), and Dermatology Life Quality Index (DLQI) were administered to obtain information on patient characteristics, AD severity, and HRQoL. The mean DLQI score for the entire sample was 11.29 ± 5.27 (range, 1–26), and 51.60% reported the disease greatly affected their lives. Bivariate analysis revealed significant differences in self-rated health measures of DLQI scores in terms of self-reported AD severity, overall health, and the EASI. In multivariable regression models adjusted for patient characteristics, the self-perceived severe AD group reported significantly higher DLQI scores (coefficient = 2.72; 95% confidence interval (CI) = 0.38–5.05; p = 0.022) than the mild group. Concurrently, we observed a substantial increase in the DLQI scores among patients with moderate and severe EASI scores (coefficient = 1.96, 95% CI = 0.08–3.92, p < 0.05 and coefficient = 4.35, 95% CI = 1.98–6.72, p < 0.001, respectively) than in those with mild EASI scores, suggesting that HRQoL was markedly influenced by greater AD severity. These findings highlight the need for a more patient-centric approach to the management of AD in order to alleviate patient suffering and, thereby, improve HRQoL.


Introduction
Atopic dermatitis (AD) is a chronic, relapsing inflammatory skin disorder characterized by frequent exacerbations of pruritic and eczematous lesions [1]. The pathogenesis of AD is complex and involves genetic susceptibility, abnormal skin barrier function, dysfunctional cell-mediated immunity, and environmental and lifestyle factors [1,2]. AD affects people of all ages, but it typically manifests in early infancy or childhood, A non-interventional, cross-sectional study was performed to assess HRQoL and associated factors in patients with AD. Participants for this study were recruited consecutively between January and April 2021 from among the patients who visited the dermatology outpatient department of the Shaheed Suhrawardy Medical College Hospital (a tertiary hospital in Dhaka, Bangladesh). The inclusion criteria were as follows: (a) patients with AD, (b) aged 16 years or older, (c) able to communicate in Bangla language, and (d) voluntarily agreeing to participate after providing written informed consent. Patients with any of the following conditions were excluded: (a) pregnant or lactating women, (b) patients with mental illness, and (c) patients having difficulty completing the surveys. We adopted the UK Working Party (UKWP) criteria to identify patients with AD [17]. This included all characteristics of the UKWP criteria, including skin itchiness during the last 12 months and three or more of the following: history of (1) skin crease involvement, (2) asthma or hay fever, (3) general dry skin during the past year, (4) visible flexural eczema, or (5) onset before the age of 2 years. Patients who met the inclusion criteria were invited to participate in the study. Patients with active dermatological diseases other than AD were excluded from the final analysis. Throughout the study, patient identification and disease severity examination were performed by one dermatologist, and this was followed by an interview by a trained enumerator.

Measurement Tools 2.2.1. Measurement of AD Severity
Disease severity was assessed using the Eczema Area and Severity Index (EASI), which is a broadly used and validated tool for measuring the severity of AD [18]. The EASI was applied to structure the observations because it is simple, easy to understand, and can be used by practitioners to normalize the baseline evaluation of the condition and track changes over time. The EASI encompasses an assessment of disease extent on a scale of 0-6 in four affected body parts (head/neck, trunk, and upper and lower extremities) and an evaluation of erythema, infiltration/papulation, excoriation, and lichenification, each scored on a scale of 0-3. The sum of these scores yields the EASI, which ranges from 0 to 72. A higher EASI score indicates extensive and severe disease conditions. In our study, we divided patients with AD into groups of equal percentiles based on their EASI scores: mild (<9), moderate (9)(10)(11)(12)(13)(14)(15), and severe (>15), as no cutoff values had been proposed for this measure.

Measurement for HRQoL
A 1-week-recall, Bangla edition of the DLQI questionnaire was administered as a self-assessment of the patients' HRQoL. The DLQI is a dermatology-specific questionnaire designed for adults (i.e., patients aged > 16 years) to evaluate any skin disorder that affects different aspects of their lives over the previous week [16,19,20]. The DLQI consists of 10 items evaluating disease impact on key aspects of patient lives and is distributed across six domains related to symptoms and feelings (Questions 1 and 2), daily activities (3 and 4), leisure (5 and 6), work and study (7), personal relationships (8 and 9), and treatment (10), as well as an overall DLQI score. Each DLQI question has four response categories ("not at all", "a little", "a lot", and "very much") that are scored on a 0-3 Likert scale. The overall DLQI is the total score of all questions and ranges from 0 to 30, with a higher score indicating increased QoL impairment. The overall DLQI score can be classified into five groups according to the influence on the patients' HRQoL: no effect (0-1), a small effect (2-5), a moderate effect (6-10), a very large effect (11)(12)(13)(14)(15)(16)(17)(18)(19)(20), and an extremely large effect (21-30).

Self-Assessed AD Severity, Overall Health, and Other Study Variables
To obtain the patients' subjective appraisal of their condition, we chose a combination of three questions dealing with previously validated "self-reported global AD severity" (how would you describe your AD: mild, moderate, or severe) [21] and "self-assessment of overall health" (how would you describe your overall health: poor, fair, good, or excellent). The demographic data collected included age, sex (male/female), educational level (≤primary, secondary, and ≥tertiary), marital status (single/married), and residence (rural/urban). We also collected information on current smoking (yes/no), family history of AD (yes/no), and allergic complications (yes/no). All the questionnaires (except for the DLQI questionnaire) were first prepared in English and then translated into Bangla.

Data Analysis
The sample size for this study was determined using G Power 3.1 software [22] based on a significance level of 0.05, a power of 0.95, and a medium-size effect of 0.15. The analysis revealed that 178 participants were required for the analysis. Patient characteristics were expressed as frequencies, percentages, means, and standard deviations. First, the internal consistency reliability of the DLQI items was examined using Cronbach's α coefficients and item-total correlations. Cronbach's α, which is a highly relevant method to measure the reliability of a scale containing multiple items, was defined for the overall instrument [16]. Cronbach's α coefficient value of not less than 0.7 is generally considered sufficient to demonstrate internal consistency [16,23]. For item-total correlation, a value of >0.3 is considered to indicate that an item is connected to the overall scale [23]. Bivariate and multivariable analyses were performed to identify the factors influencing HRQoL among patients with AD. First, a bivariate analysis was conducted to explore the differences across the patients' general characteristics using domain-specific DLQI and overall DLQI scores. Non-parametric approaches, the Mann-Whitney U-test and the Kruskal-Wallis test, were used because the variables had a non-normal distribution according to the Shapiro-Wilk test. Subsequently, posthoc F tests were performed to detect any potential differences between the groups.
Second, multiple linear regression analyses were adopted to determine the association between illustrative variables: domain-specific DLQI, overall DLQI, and patient characteristics. Individual variables that showed statistical significance (p < 0.05) at the bivariate level were included in the regression model. Domain-specific DLQI (symptoms and feelings, daily activities, leisure, work and school, personal relationships, and treatment) and overall DLQI scores were used separately as dependent variables in the model, and the potential confounders adjusted for were marital status, family history of AD, self-reported global AD severity, overall health ratings, and EASI scores. We included the variance inflation factor (VIF) in each regression model to check for multicollinearity; however, none of the variables exceeded a VIF > 2.0. The beta coefficient was also determined to evaluate the strength of the association with a 95% confidence interval (CI). The significance level was set to p < 0.05. Statistical analyses were performed using IBM SPSS Statistics for Windows/Macintosh, Version 25.0 (IBM Corp., Armonk, NY, USA).

Patient Characteristics
Of the 198 patients enrolled, 14 were excluded because they had active skin diseases other than AD, thus leaving 184 to be included in the final analysis ( Figure 1). Table 1 presents the descriptive characteristics of the patients. The participants included 83 men (45.10%) and 101 women (54.90%) with a mean age (±standard deviation) of 33.46 ± 15.44 years (range, 16-70 years), and the highest proportion of patients was in the age group below 30 years (54.30%). In total, 52.20% of patients had completed tertiary education, and 57.60% were married. Moreover, 97 (52.70%) patients reported having a family history of AD, and 98 (53.30%) had allergy complications. In terms of self-reported global AD severity, 21 (11.40%) patients reported having mild AD, 51 (27.70%) as having moderate AD, and 112 (60.90%) as having severe AD. The overall health rating of the patients was poor in 37 (20.10%), fair in 35 (19.00%), good in 98 (53.30%), and excellent in 14 (7.60%). The distribution of disease severity was expressed on the basis of the cutoff value of the EASI score. The severity of AD was mild, moderate, and severe in 33 (17.90%), 97 (52.70%), and 54 (29.40%) patients, respectively. EASI score. The severity of AD was mild, moderate, and severe in 33 (17.90%), 97 (52.70%), and 54 (29.40%) patients, respectively.

Descriptive Summary of DLQI Scale Scores, Internal Consistency, and Overall HRQoL
The descriptive summary and internal consistency reliability of the DLQI items are presented in Table 2. The internal reliability coefficient (measured using Cronbach's α) for the DLQI was 0.75. As shown in Table 2, Cronbach's α ranged between 0.71 and 0.76 for all 10 items of the DLQI questionnaire, indicating good internal consistency and no redundant items. The item-total correlation for all the 10 items exhibited good correlation (ranging between 0.23 and 0.63), except for Item 9. The mean overall DLQI score for the whole sample was 11.29 ± 5.27 (range, . Of the 10 DLQI scales, the highest score was recorded for Question 1 (itchy and painful; 2.15 ± 1.08), followed by Question 2 (embarrassment; 1.89 ± 1.06). Figure S1 represents the distribution of patient responses on the DLQI items. Based on the overall DLQI score classification, only 2.7% of patients reported AD having "no effect on their life", while 28.80% and 51.60% of patients reported AD having "mild" and "very large effect on their life", respectively (Table 3).

HRQoL According to Patient Demographic Characteristics
The association between domain-specific and overall DLQI and patient demographic characteristics at the bivariate level is shown in Table 4. Significant differences in self-rated health measures of domain-specific DLQI and overall DLQI scores were found in relation to marital status and family history of AD. Scores of "daily activities", "leisure", "personal relationship", and overall DLQI were significantly higher in the married group than in the single group. Patients who had a family history of AD reported higher scores in "symptoms and feelings", "daily activities", "leisure", "personal relationship", and overall DLQI.  3.4. HRQoL According to Patient Self-Perceived AD Severity, Overall Health Rating, and the EASI Bivariate association between the patients' HRQoL and self-perceived AD severity, overall health rating, and the EASI are shown in Table 5 and Figure 2. A significant difference was observed in "symptoms and feelings", "daily activities", and "personal relationship" in relation to self-perceived AD severity and the EASI. Likewise, the overall DLQI scores were significantly higher (p < 0.01) in the self-perceived moderate and severe AD groups than in the mild AD group (Figure 2a). Those with poor and fair self-perceived overall health rated significantly higher DLQI scores (p < 0.01 and p < 0.05, respectively) than did those with excellent overall health (Figure 2b). We also observed a significant increase in the overall DLQI score in the moderate and severe AD groups (p < 0.001 and p < 0.01, respectively) than in the mild AD group (Figure 2c). Values of domain-specific and overall DLQI expressed as mean ± SD. DLQI: Dermatology Life Quality Index; SD: standard deviation; AD: atopic dermatitis; EASI: Eczema Area and Severity Index; p-value calculated using the Mann-Whitney Utest or Kruskal-Wallis test; values in bold are significant. a EASI categorized as: scores <9 (mild), 9-15 (moderate), and >15 (severe).

Factors Influencing HRQoL
The models used to estimate the factors influencing domain-specific DLQI are presented in Table 6. "Leisure"-related QoL was significantly associated with a "family history of AD" (p < 0.01). Patients with "self-perceived severe AD" were more likely to have problems with QoL related to "symptoms and feelings" (p < 0.01) and "work and school" (p < 0.05). However, we found that all the domain-specific DLQI scales (except "work and school") were significantly associated with severe EASI scores, indicating that severe EASI groups are more likely to experience problems with QoL associated with "symptoms and feelings", "personal relationships", and "treatment". In Table 7, the regression model revealed that the patients' HRQoL was considerably influenced by a "family history of AD" (coefficient = 1.49; 95% CI = 0.02-2.97; p < 0.047). Then again, in "self-reported global AD severity", the severe AD group showed a significant association with the overall DLQI (coefficient = 2.72; 95% CI = 0.38-5.05; p < 0.05), indicating poorer HRQoL than in the mild AD group. Concurrently, a significant increase in DLQI scores was observed among patients with moderate and severe EASI scores (coefficient = 1.96, 95% CI = 0.08-3.92, p < 0.05 and coefficient = 4.35, 95% CI = 1.98-6.72, p < 0.001, respectively) than among those with mild EASI scores, suggesting a marked impairment in HRQoL with disease severity. However, no significant association with HRQoL was observed for sociodemographic parameters and overall health ratings.

Discussion
To our knowledge, this is the first study in Bangladesh to assess HRQoL and associated factors in adults with AD, including a test to investigate the internal consistency reliability of the DLQI items. Internal consistency reliability measures the homogeneity of multiple items and is considered a test of the reproducibility of an instrument [16]. For instance, when a patient's health status remains stable over time, the results of the instruments should also remain stable [9]. In our study, we found that the internal consistency reliability of the DLQI items was good, as judged using Cronbach's α coefficients and item-total correlations. These findings are similar to those obtained in the original validation study and studies using the DLQI in other languages [16,19]. Therefore, we suggest that the Bangla version of the DLQI may serve as a potential instrument for AD-related health status measurement in Bangladesh and may contribute to cross-cultural comparison. In addition, the DLQI offers advantages such as brevity, low burden to participants and data collectors, and ease of interpretation. Furthermore, the implementation of the DLQI would help clinicians address specific complications in the patients and categorize which aspects of the patients' lives are most severely affected [20].
We observed that poorer HRQoL was associated with the DLQI scores; half of the participants reported having very large effects on their lives, and these outcomes were mostly influenced by greater AD severity. Among the participants, the mean DLQI score was 11.29, which was relatively higher than that reported in several previous studies in HICs [21,[24][25][26]. Notably, a cross-sectional population-based study of 602 US adults reported a mean DLQI score of 4.9 [26]. In contrast, Vakharia et al. found a mean DLQI score of 10.7 among 265 US adults, while Holm et al. and Sanchez-Perez et al. found mean DLQI scores of 9.79 and 7.8, respectively, among Danish and Spanish patients [21,24,25]. However, a recent review of 27 studies in Asia found that the DLQI values ranged between 4.8 and 12.0 [27]. Similarly, another review of 32 studies across the US, Europe, and some Asian countries suggested that the mean DLQI scores varied between 4.9 and 20 [7]. The same review addressed the difficulty in interpreting the overall score because of its variability across studies and the lack of formal reference values or population normal scores for formal comparisons. Therefore, more in-depth qualitative research is needed to confirm these findings.
A substantial number of reports support the role of disease severity as an important risk factor for HRQoL impairment in patients with AD [5][6][7][8]27]. In our study, we examined the influence of AD on HRQoL in association with disease severity, measured using both the clinical objective scoring system EASI and self-rated global AD severity. Both bivariate and multivariable regression analyses showed that patients with moderate and severe AD reported worse HRQoL than those with mild AD. This is consistent with the findings of previous reports that support the hypothesis that HRQoL decreases as disease severity increases [28,29]. In a systematic review of 15 studies in Asia, 14 studies showed a significant association between disease severity and HRQoL impairment [27]. In another review, 19 out of 20 studies showed that increased disease severity was significantly associated with poorer HRQoL [7]. The effects of AD on impaired HRQoL in our study seem mostly related to itchiness and embarrassment as these parameters were the most commonly recorded in the DLQI. Notably, chronic pruritus and eczematous lesions are the most common symptoms in AD that may affect the patient's mood and lead to embarrassment, thereby impacting the patient's psychological well-being. In contrast, greater severity was associated with greater impairment in several dimensions of the DLQI, namely, symptoms and feelings, daily activities, leisure, personal relationships, and treatment, which is consistent with the findings of previous research [29,30]. Evidence also suggests that personal relationships were the least affected dimension of HRQoL [7]; however, we found that patients aged between 30 and 50 years experienced more problems in personal relationships than did those in other age groups. Moreover, we noticed that having a family history of AD negatively influenced HRQoL, especially leisure-related QoL. As AD is a restrictive disease that interferes with daily activities, social and interpersonal relationships would obviously be affected in patients. Furthermore, in LMICs, limited access to treatment, disease management, lack of formal diagnosis, and low socioeconomic status, which possibly result in poor follow-up in clinics, may contribute to impaired HRQoL in patients with AD [11,12,15].
Several studies showed no clear association between the patients' demographic characteristics and HRQoL impairment [27,31]. Our analysis also did not reveal such an association between demographic characteristics and HRQoL. However, Holm et al. found more impaired HRQoL in female patients than in male patients [24]. Therefore, further studies are required to clarify these relationships. Self-perceived poor overall health was associated with AD in a study by Silverberg et al. [26]. In our study, the bivariate analysis showed an association between self-perceived overall health and HRQoL in AD; however, in the regression model adjusted for potential confounders, this association was diminished. One explanation for this observation may be that the participants in our study visited a tertiary hospital because of the main complaint of skin problems. Therefore, disease severity seems to be the main factor that influences HRQoL in Bangladeshi adult patients with AD.
Several limitations of this study warrant consideration. First, the outcomes of a crosssectional study cannot be taken as evidence of causal relationships. Second, the single-institution study design limits the generalizability of the results. Third, we did not collect information on the age of AD onset, current treatment therapy, sleep disorders, and HRQoL using generic questionnaires. Lastly, we did not include healthy subjects as a control group; therefore, our observations need further validation. Nevertheless, to our knowledge, this study is the first of its kind in Bangladesh to document HRQoL among adults with AD. In addition, we used the objective scoring system EASI to evaluate disease severity. The EASI allows for individual items to be measured, either separately or in combination, thereby providing a more comprehensive evaluation and acting as a reference for clinical disease severity in patients. Further research on a larger patient cohort, including more medical information and utilizing both generic and dermatology-specific QoL questionnaires, is required to better quantify the effects of AD and ensure effective treatment.

Conclusions
In conclusion, this study demonstrates that Bangladeshi adults with AD experience substantially poorer HRQoL and that the degree of impairment is markedly influenced by greater AD severity. Most of the patients experienced the effects of "symptoms and feelings", "daily activities", "leisure", "personal relationships", and "treatment" on their QoL. In addition, employing the DLQI instrument seemed beneficial in identifying vulnerable patients for whom appropriate treatment strategies have to be designed and for whom the therapeutic response has to be measured. Therefore, we recommend that dermatologists be aware of the clinical characteristics and negative psychosocial impacts of AD on the HRQoL of patients with this condition.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/ijerph182111593/s1, Figure S1: Distribution of patients' responses on the DLQI items (n = 184).  Informed Consent Statement: Informed consent was obtained from each participant before enrollment and after explaining the research objectives and consequences and providing an assurance of confidentiality.