Development and validation of the Topical Steroid and Moisturizer Adherence Scale

Abstract

Full text

To the Editor: While dermatologic outcomes are often compromised by nonadherence to prescribed topical therapies,1 the ability to measure and track topical nonadherence is limited by the lack of validated instruments. Existing measures such as tube weighing2 and smart-cap trackers3 are costly and limited by variability in treatment requirements that vary with disease severity. We developed and prospectively validated a 7-item tool—Topical Steroid and Moisturizer Adherence Scale (TSMAS)—a simple and practical patient-reported measure to assess adherence to topical corticosteroids (TCS) and moisturizers.

Items were developed based on literature search of 22 articles and expert opinion, comprising questions on TCS adherence (frequency, quantity, potency, and overall use) and moisturizer adherence (frequency, quantity, and overall use). Items employed Likert-type response options with a 30-day recall period.

Face validity and cognitive debriefing were performed in 5 patients, followed by validation in 119 dermatology outpatients from the National University Hospital, Singapore, between August 2022 and May 2024. Patients aged 16 to 94 years (median 42 years), prescribed TCS and moisturizers for a chronic inflammatory skin disease completed the questionnaire at baseline and 1 month (Table I). To support construct validity, the questionnaire included patient-reported outcomes and clinician assessments (Table II). Standard psychometric methods were used to assess internal consistency, test–retest reliability, construct validity, and responsiveness.Table IBaseline characteristics of participants (N = 119)CharacteristicValueAge, mean (SD)44.5 (19.2) ySex, n (%) Male62 (52.1%) Female57 (47.9%)Ethnicity, n (%) Chinese82 (68.9%) Malay18 (15.1%) Indian9 (7.6%) Other10 (8.4%)Highest education level, n (%) Primary school2 (1.7%) Secondary school22 (18.5%) Junior college/ITE43 (36.1%) Bachelor’s degree41 (34.5%) Master’s/Doctorate11 (9.2%)Disease condition, n (%) Eczema∗92 (77.3%) Prurigo nodularis1 (0.8%) Autoimmune blistering disease4 (3.4%) Psoriasis20 (16.8%) Others (Jessner’s lymphocytic infiltrate, Cutaneous lupus)2 (1.7%)∗Includes all eczema variants (Atopic dermatitis, hand and feet eczema, discoid eczema, asteatotic eczema).Table IIOverview of questionnaire itemsConstructsQuestionsAnswer optionsTopical Steroid and Moisturizer Adherence Scale (TSMAS) Adherence to Topical Steroids (TCS)Compared to what I’ve been prescribed and asked to use, I amAlways adherent; Sometimes adherent; Seldom adherent; Nonadherent/I don’t useCompared to how often I am supposed to apply the steroid creams prescribed to me, I use itMuch more frequently; Slightly more frequently; As prescribed; Slightly less frequently; Much less frequently; I don’t useCompared to how much I am supposed to apply the steroid creams prescribed to me, I use aMuch larger amount; Slightly larger amount; As prescribed; Slightly smaller amount; Much smaller amount; I don’t useCompared to what I was prescribed, most of the time, I useMore of the stronger steroid cream; As prescribed; More of the weaker steroid cream; I don’t use; Not relevant (single strength) Adherence to MoisturizersCompared to what I’ve been prescribed and asked to use, I amAlways adherent; Sometimes adherent; Seldom adherent; Nonadherent/I don’t useHow frequently do you use moisturizers?Twice daily or more; Once daily; 2-4 × weekly; Weekly or less; I don’t useHow much moisturizers do you apply?Large and thick amount; Moderate amount; Small and thin amount; I don’t useBaseline patient reported tools for evaluation of psychometric properties QoL impairment (Skindex mini) Anxiety (GAD-7) Social desirability Self-reported severity Self-reported body surface area (BSA) Desire for improvement Steroid adherence barriers Moisturizer adherence barriersConstructs assessed at 1 mo follow up Adherence measure—steroids and moisturizers Self-reported severity Self-reported body surface area (BSA)Information provided by the physician at baseline BSA Investigator’s global assessment (IGA) Patient’s knowledge of steroid cream (name and potency) Patient’s technique of application Doctor’s estimated quantity of TCS usage (fraction of used vs should have used)

Item-level ceiling effects were observed, with >50% indicating full adherence for questions on prescribed steroid quantity and potency, overall moisturizer adherence and frequency of application. However, the domain-level summed scores did not exhibit ceiling effects. Internal consistency was good for TCS (α = 0.86) and modest for moisturizer-related items (α = 0.66). Confirmatory factor analysis (CFA) showed acceptable fit for a 2-factor model with a Comparative Fit Index (CFI) of 0.93, Tucker-Lewis Index (TLI) of 0.89 and Standardized Root Mean Square Residual (SRMR) of 0.078, though the Root Mean Square Error of Approximation (RMSEA) of 0.116 was suboptimal suggesting room for further optimization (detailed results in Supplementary File, available via Mendeley at https://data.mendeley.com/datasets/ygvc7zmj85/1).

Convergent validity was demonstrated with physician-rated adherence (ρ = 0.34), perceived benefits (ρ = 0.26), and desire for skin improvement (ρ = 0.38). In contrast, weak or non-significant correlations with anxiety (ρ = 0.18), quality of life (ρ = 0.02) and education (ρ = −0.04) supported discriminant validity. There was no correlation between adherence and social desirability, suggesting no interviewer or social desirability bias.4

Test–retest reliability was moderate for the TCS domain (intraclass correlation coefficient: 0.762) and for the moisturizer domain (0.569). Responsiveness was moderate, with standardized mean response (SMR) of 0.585 and −0.596 for improved and reduced adherence in the steroid domain, and 0.360 and −0.385, respectively, in the moisturizer domain.

The TSMAS demonstrates acceptable reliability and validity for measuring adherence. It offers a simple, low-cost way to detect nonadherence and guide management. Study limitations include a non-normal item distribution, a small item pool and the absence of a gold standard criterion measure precluding criterion validity evaluation. A ceiling effect also limited discrimination at higher adherence levels. Future work can explore refinements to response options (see proposed questionnaire in Supplementary File, available via Mendeley at https://data.mendeley.com/datasets/ygvc7zmj85/1), optimization through item refinement, model re-specification, and evaluation across broader populations to establish clinically meaningful thresholds.5

Conflicts of interest

None disclosed.