Study design and setting

This cross-sectional validation study was conducted at the Tirana Regional Hospital Center ‘Shefqet Ndroqi’, Albania’s national referral center for pulmonary diseases, from March 2024 to March 2025. All study procedures were approved by the Ethics Committee of the University of Medicine, Tirana (approval number 2096/1 Prot., dated 01.08.2024), and written informed consent was obtained from every participant.

Participants

Eligible participants were adult outpatients (≥18 years) with a confirmed diagnosis of asthma or COPD, verified by spirometry and documented in accordance with GOLD (Global Initiative for Chronic Obstructive Lung Disease, 2024) and GINA (Global Initiative for Asthma, 2024) criteria^4,5. Only patients who were currently prescribed at least one inhaler and fluent in Albanian were included. Patients with cognitive impairment, inability to consent, or an acute exacerbation in the previous two weeks were excluded.

Study instrument

The TAI is a 12-item self-administered instrument developed by Plaza et al.¹⁰ to assess medication adherence in patients with asthma and COPD. The first ten items (Q1–Q10) are completed by the patient and capture adherence behavior, including missed doses, intentional dose reduction, lifestyle interference, and beliefs about treatment necessity. Each item is rated on a 5-point Likert scale (1=worst adherence/all week to 5=best adherence/never), yielding a subscale score from 10 to 50. The remaining two items (Q11–Q12) are clinician-administered and assess inhaler technique and patient knowledge of the prescribed regimen, each rated from 1 to 2 (subscale score 2 to 4). The combined 12-item total ranges from 12 to 54, with established cutoffs of ≥50 indicating high adherence, 46–49 intermediate adherence, and ≤45 low adherence¹⁰. The full Albanian version is provided in the Supplementary file.

Following the methodology of the original Spanish validation¹⁰ and subsequent cross-cultural adaptations^12-15, the present psychometric analyses were performed only on the ten self-report items (Q1–Q10). The clinician-rated items (Q11–Q12) were collected during routine clinical assessment but were not included in reliability or factor analyses, to preserve methodological comparability with previous TAI validation studies.

Translation and cultural adaptation

The TAI-12 was translated into Albanian from the established English version, as English is the working language of the international validation literature^12-15 and was the version available to our bilingual translators. Permission for academic use of the TAI was obtained through the BiblioPRO patient-reported outcome measures registry, the authorized distribution platform for the questionnaire²¹. Forward translation was performed independently by two native Albanian speakers with clinical training (a clinical pharmacist with 15 years of clinical experience and a medical resident, with 5 years of experience), both clinically informed and with prior PROM translation experience. The two forward versions were reconciled by consensus. Back-translation into English was performed independently by a pulmonology resident who was not involved in the forward translation and who was blinded to the original English instrument. The expert panel, comprising the lead author (IA, clinical pharmacist with 4 years of experience in respiratory pharmacotherapy) and a senior pulmonologist (DhA, with 32 years of clinical experience in COPD and asthma management at the national pulmonary referral center), reviewed the reconciled translation for linguistic clarity, semantic equivalence, and cultural relevance. The final draft was pilot-tested for clarity and feasibility with five healthcare professionals from the Tirana Regional Hospital Center ‘Shefqet Ndroqi’ (two senior pulmonologists with more than 20 years of clinical experience, two pulmonology residents in their final year of training, and one respiratory nurse with eight years of clinical experience). Cognitive debriefing within the target patient population was not performed, which is acknowledged as a limitation of the present translation methodology.

Recruitment and data collection

Trained pulmonologists and residents administered the Albanian TAI during routine outpatient visits. Demographic and clinical data (age, sex, education level, occupation, years since diagnosis, COPD stage according to GOLD 2024 A–E groups; asthma control as mild, moderate, or severe based on the GINA 2024 treatment steps required to achieve control, FEV1 %, number of prescribed inhalers) were collected via patient interview and medical record review. Data were entered into a secure, anonymized electronic database.

Statistical analysis

Continuous variables are summarized as mean and standard deviation (or median and range where distributions deviated from normality), and categorical variables as frequencies and percentages. Internal consistency was estimated using Cronbach’s alpha and item-rest correlations. Because this represents the first validation of the TAI in the Albanian cultural context, we used exploratory factor analysis (EFA) rather than confirmatory factor analysis to allow the underlying factor structure to emerge without imposing a pre-specified model. Suitability of the data for factor analysis was assessed using the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity. Structural validity was evaluated through principal axis factor analysis with oblimin rotation, retaining factors with eigenvalues >1, and examining the scree plot for dimensionality.

Known-groups validity was assessed by comparing TAI scores across COPD and asthma severity categories using Welch’s ANOVA with Games–Howell post hoc tests, given violation of the homogeneity-of-variance assumption.

Criterion validity was assessed by correlating TAI scores with FEV1 % predicted, calculated using the Global Lung Function Initiative (GLI) 2012 equations, and with the number of prescribed inhalers, using Pearson’s r. Analyses were performed using Jamovi (version 2.7.12) together with R (version 4.5) and the psych package^22-24. A two-sided p<0.05 was considered statistically significant.

Participant characteristics

A total of 326 patients participated, with a range of COPD and asthma severity represented in the sample. Mean age was 66.6 years (SD=11.9), 240 (73.6%) were male, and the sample was approximately evenly distributed between urban (n=170; 52.1%) and rural (n=156; 47.9%) residence. Most patients had completed secondary school (n=222; 68.1%), with smaller proportions in primary or 9-year schooling (n=66; 20.2%), university (n=30; 9.2%), or no formal schooling (n=8; 2.5%), reflecting national trends of education level in Albania²⁵. Mean FEV1 was 57.7 ± 20.5% predicted. The sample comprised 210 patients with COPD (64.4%; GOLD Group A, n=15; Group B, n=81; Group E, n=114) and 116 with asthma (35.6%; mild, n=38; moderate, n=43; severe, n=35). Patients were prescribed one (n=42; 12.9%), two (n=156; 47.9%), or three (n=128; 39.3%) inhalers. Full sociodemographic and clinical characteristics are presented in Table 1.

TAI item characteristics

All ten TAI self-report items were completed by every participant, with no missing data. Item means ranged from 2.7 to 4.2, indicating adequate variability and absence of floor or ceiling effects. Table 2 summarizes item-level metrics, including mean, SD, skewness, and kurtosis, demonstrating distributions suitable for psychometric analysis.

Cronbach’s alpha for the 10-item scale was 0.64. Item-rest correlations were highest for items 1 to 3 (r=0.507, 0.679, and 0.554, respectively) and lowest for items 5, 6, and 8 (r=0.009, 0.056, and 0.140, respectively) (Table 3).

Exploratory factor analysis supported a unidimensional structure: Factor 1 (eigenvalue 2.30) accounted for 23 % of total variance. Items 1–3 exhibited the highest loadings, while items 5–10 showed low loadings and high uniqueness (Table 3). The Kaiser–Meyer–Olkin statistic was 0.74, with Bartlett’s test confirming suitability for factor analysis (χ²=587, p<0.001).

Regarding known-groups validity, TAI total scores differed significantly across COPD stages (Welch’s F=18.7, p<0.001). Patients with advanced COPD (stage E) reported the highest average adherence (mean=37.7, SD=4.18), followed by those at moderate (mean=34.5, SD=4.33) and early stages (mean=32.9, SD=4.12) (Figure 1).

Figure 1

Total TAI score by GOLD A–E group in adults with COPD, cross-sectional validation study, Tirana, Albania, March 2024–March 2025 (N=210)

Similarly, TAI total scores differed significantly across asthma severity groups (Welch’s F=4.83, p=0.011). Post hoc comparisons using the Games-Howell test revealed that patients with severe asthma reported significantly higher adherence (mean=36.7, SD=4.42) compared to those with mild asthma (mean=33.5, SD=4.52, p=0.008). No significant differences were found between moderate asthma (mean=34.7, SD=6.08) and the other stages (Figure 2).

Figure 2

Total TAI score by GINA severity category in adults with asthma, cross-sectional validation study, Tirana, Albania, March 2024–March 2025 (N=116)

Correlation analyses revealed an inverse association between TAI score and FEV1 % predicted (r= -0.298, p<0.001) and a positive correlation with the number of inhalers prescribed (r=0.209, p<0.001).

Limitations

The present study has several limitations. Recruitment was conducted at a single tertiary referral center using consecutive convenience-based sampling, which may have introduced selection bias toward more complex diseases and may limit generalizability to primary care or non-specialist settings in Albania. Internal consistency was modest, and limited variance was explained by the single extracted factor, with several items contributing weakly. Cognitive debriefing within the target patient population was not performed; pilot testing was limited to healthcare professionals, which may have affected items addressing culturally loaded concepts.