Empowering Adherence Assessment: Translating and Validating the 15-STARS Questionnaire for Use in Taiwanese Community Pharmacies Authors: Yung-Hsuan Chang, Yen-Ming Huang, Yu-Meng Yang, Tzu Wang, Hsun-Yu Chan, Chia-chen Yang, Isabelle Arnet, Christiane Eickhoff, Yunn-Fang Ho Published: 2026 ABSTRACT BackgroundMedication adherence is a key determinant of treatment effectiveness, yet few brief, freely available, and culturally appropriate adherence assessment tools are available in Traditional Chinese for use in routine community pharmacy practice. The 15-item Screening Tool for AdheRence to medicineS (15-STARS) is a pharmacy-based questionnaire designed to identify medication non-adherence and its underlying barriers.ObjectiveTo translate, culturally adapt, and validate the 15-STARS in Traditional Chinese and evaluate its utility for screening medication non-adherence and identifying barriers to adherence among patients with type 2 diabetes (T2D) in Taiwanese community pharmacies.MethodsThis cross-sectional study comprised two sequential phases: translation and cross-cultural adaptation, followed by field testing. Translation involved forward-backward translation and cognitive debriefing with pharmacists and patients. The Traditional Chinese 15-STARS was administered to adults with T2D recruited from six community pharmacies in Taiwan. Psychometric evaluation included content validity, item-level criterion validity against the Traditional Chinese Adherence to Refills and Medications Scale (ChARMS-T), and internal consistency assessed using McDonald’s omega (ω). Associations among adherence, medication-related factors, and glycemic control were examined.ResultsTen pharmacists and ten patients participated in cognitive debriefing, and 215 patients completed field testing. Content validity was excellent, with item-level and scale-level content validity indices of 1.00 after revision. Items 6–9 demonstrated good internal consistency (ω = 0.831), and selected 15-STARS items were significantly correlated with corresponding ChARMS-T measures. Among the participants, 53.0% of participants reported no missed medication doses during the previous two weeks, whereas 47.0% reported missing at least one dose. The most common reasons for non-adherence were forgetfulness (33.0%), concerns about taking too many medications (18.6%), and concerns about side effects (15.3%). Greater regimen complexity and higher dosing frequency were associated with lower adherence and poorer glycemic control.ConclusionThe Traditional Chinese 15-STARS demonstrated satisfactory content validity, evidence of criterion validity for selected items, and internal consistency among conceptually related belief-based items in Taiwanese community pharmacies. The instrument may serve as a practical screening tool for identifying patients at risk of non-adherence and exploring adherence-related barriers, thereby supporting targeted pharmacist-led interventions. Introduction Medication adherence refers to the extent to which patients take medications as prescribed by healthcare professionals.1 It is commonly conceptualized as a dynamic process comprising three components: initiation, implementation, and discontinuation.2 As famously stated by former US. Surgeon General C. Everett Koop, “Drugs do not work in patients who do not take them”.3 Although pharmacotherapy is central to the management of chronic diseases, its effectiveness ultimately depends on patients’ ability and willingness to take medications as prescribed. The World Health Organization estimates that approximately 50% of individuals with chronic conditions do not adhere to their prescribed medication regimens.4 Poor medication adherence has been associated with adverse clinical outcomes, increased hospitalization rates, and greater healthcare expenditures.5 Accordingly, healthcare professionals must proactively identify suboptimal medication-taking behaviors and tailor treatment strategies to patients’ needs.6 A wide range of approaches has been developed to assess medication adherence; however, no single method can be regarded as a definitive standard. Direct measures, such as quantifying drug or metabolite concentrations in blood or urine, provide objective evidence of medication use but are invasive and feasible only in selected clinical settings.7 Indirect objective approaches include the analysis of administrative claims data, such as Taiwan’s National Health Insurance database, to estimate adherence based on prescription refill records. These methods rely on the assumption that medication refills reflect actual medication-taking behavior, which may not always be the case.8 Other objective measures include directly observed therapy,1 electronic medication packaging devices that record dosing events, and pill counts that estimate doses taken between visits or refills.7 In contrast, subjective measures rely on assessments of medication-taking behavior reported by patients or evaluated by healthcare professionals, most commonly through structured questionnaires.9 While objective measures offer quantifiable data, they often provide limited insight into patients’ beliefs, motivations, and decision-making processes related to medication use. Subjective measures can help address these gaps by capturing patients’ perspectives and perceived barriers. Conversely, objective measures may serve as external validation criteria for subjective tools, helping to mitigate potential self-reporting and social desirability bias.7 Previous studies have shown that self-reported adherence questionnaires demonstrate moderate to high concordance with electronic monitoring devices.10 Improving medication-taking behavior requires not only identifying non-adherence but also understanding the underlying reasons for patients’ decisions.11 Factors contributing to non-adherence span multiple levels, including limited access to healthcare services, suboptimal patient-provider communication, regimen complexity, and individual-level factors such as medication self-efficacy and beliefs about medicines.12 In routine clinical practice, medication adherence is often assessed using a single, general question about whether patients take their medications as prescribed. However, patients may be reluctant to disclose non-adherence due to concerns about negative reactions from healthcare providers.13 Self-report questionnaires offer a less intrusive and more comprehensive approach by assessing multiple dimensions of the medication-taking process through structured, multi-item instruments.14 These tools also facilitate the identification of diverse adherence barriers, including financial constraints, limited time for clinical consultations, disease-related challenges, adverse effects, and patient beliefs. The growing recognition of the value of self-report adherence measures has led to the development of numerous questionnaires. For use in routine clinical settings, particularly community pharmacies, such tools must be affordable, time-efficient, and easy to administer.15 Moreover, questionnaires must be linguistically and culturally validated to ensure accurate and reliable assessment of medication-taking behavior within local contexts.16 In Taiwan, relatively few medication adherence questionnaires are available in Traditional Chinese. Existing instruments may be limited by excessive length, high licensing costs, or lack of culturally adapted versions suitable for routine clinical practice. For example, the Hill-Bone Compliance Scale for hypertension and the Medication Adherence Rating Scale for psychosis have not been translated into Traditional Chinese.17 Other instruments face practical barriers. The Traditional Chinese version of the Medication Adherence Reasons Scale comprises 24 items, which may limit its feasibility in routine practice.18 In addition, the Morisky Medication Adherence Scale requires licensing fees, which may restrict its widespread use.19 Community pharmacies represent a valuable setting for medication adherence assessment, as patients receive both medications and professional counseling during routine visits.20 To be feasible in this context, adherence questionnaires must be cost-effective and completable within approximately 5-10 minutes. The 15-item Screening Tool for AdheRence to medicineS (15-STARS) is a practical option, as it is free of charge and requires only 3-5 minutes to complete.21 The instrument screens medication-taking behavior, identifies patient-related barriers to adherence, and assesses patients’ need for additional professional support. The 15-STARS has been validated in English, French, and German populations,21 but a Traditional Chinese version suitable for use in Taiwan is currently unavailable. In recent years, several Traditional Chinese instruments related to medication use have been developed and validated in Taiwan, including the Adherence to Refills and Medications Scale (ChARMS-T) for assessing medication-taking and refill adherence,8 the Beliefs about Medicines Questionnaire--Specific for evaluating medication beliefs,22 and the Medication Understanding and Use Self-Efficacy Scale for measuring self-efficacy in medication adherence.23 While these instruments provide valuable insights into specific aspects of medication use, each focuses on a specific construct and may not capture the range of adherence behaviors and barriers relevant to routine medication management. In contrast, the 15-STARS was specifically designed as a multidimensional screening tool that evaluates adherence behavior, practical medication-use difficulties, reasons for non-adherence, medication discontinuation behaviors, and patients’ need for professional support. To address this gap, the present study aimed to translate and culturally adapt the English version of the 15-STARS into Traditional Chinese and to evaluate its psychometric properties in community pharmacy settings in Taiwan. We also sought to characterize medication adherence and adherence-related barriers among individuals with type 2 diabetes (T2D) and explore the instrument’s potential utility as a practical screening tool to support patient-centered adherence services in routine community pharmacy practice. Methods Study Overview This cross-sectional study comprised two sequential phases: (1) translation and cross-cultural adaptation and (2) field-testing. Both phases were conducted in collaboration with community pharmacies that partnered with National Taiwan University to support pharmacy student experiential training. During the translation phase, established forward-backward translation and cultural adaptation procedures were followed to develop a Traditional Chinese version of the 15-STARS in accordance with recommended guidelines.24,25 In the field-testing phase, the translated instrument was administered to patients recruited from participating community pharmacies to evaluate its psychometric properties, including validity and reliability. This phase also assessed patients’ medication-taking behaviors and adherence-related barriers. An overview of the study design and procedures is presented in Figure 1. The translation and cross-cultural adaptation phase was conducted between February and July 2023, whereas recruitment for the field-testing phase took place over a nine-month period, from September 2023 to May 2024. Figure 1Flowchart of the study design and procedures across study phases.A flowchart outlines study phases: Forward Translation, Backward Translation, Cognitive Debriefing (Pharmacists and Patients), Survey Implementation and Data Analysis. Forward Translation involved three independent translations from English to Traditional Chinese, followed by a meeting to finalize the version. Backward Translation translated back to English, refining the Mandarin version. In Cognitive Debriefing (Pharmacists), ten pharmacists assessed comprehensiveness and style, with adjustments made. Cognitive Debriefing (Patients) had ten patients evaluate clarity and ease, with necessary changes. Survey Implementation included interviews with 215 diabetes patients at six pharmacies. Data Analysis assessed reliability, validity, and reasons for medication non-adherence. The phases were divided into translation and cross-cultural adaptation and field-testing.A flowchart of study phases: translation, adaptation and field-testing procedures. Flowchart of the study design and procedures across study phases. Translation and Cross-Cultural Adaptation Phase Forward Translation Three native Chinese speakers with high proficiency in English independently translated the original questionnaire into Traditional Chinese. All translators were familiar with pharmacy practice. In accordance with recommended guidelines,25 each translator was instructed to prioritize commonly used healthcare language over technical or professional jargon to enhance comprehensibility for the general population. The three forward translations were subsequently reviewed in a consensus meeting, during which discrepancies were discussed and resolved to produce a single reconciled version. Cultural adaptation was conducted concurrently to ensure conceptual equivalence and contextual appropriateness. Backward Translation and Review The reconciled forward translation was then translated back into English by an independent bilingual translator who had no professional background in pharmacy and was blinded to the original questionnaire and its underlying constructs. A review meeting involving all translators was held to compare the back-translated version with the original English version. Discrepancies were discussed to identify potential issues related to semantic or conceptual equivalence. Based on this review, revisions were made to the Traditional Chinese version to improve accuracy and clarity. Cognitive Debriefing with Pharmacists Cognitive debriefing was conducted with ten pharmacists (five men and five women) to assess item relevance, clarity, and linguistic appropriateness. Pharmacists rated each item and sentence using a 6-point Likert scale (1 = very poor, 6 = very good) provided by the Pharmaceutical Care Network Europe (PCNE). Feedback on wording and phrasing was also collected. Items receiving a score below 4 (defined as sufficient) were discussed by the research team, and revisions were made accordingly. Cognitive Debriefing with Patients The revised version was subsequently evaluated through cognitive debriefing with ten patients. Participants assessed item clarity and ease of understanding using a 4-point Likert scale (1 = very poor, 4 = very good), also provided by the PCNE. Participants were encouraged to provide feedback on how to improve comprehensibility. Convenience sampling was used, with efforts made to ensure diversity in age and educational background and equal representation of men and women. Items scoring below 3 (good) were reviewed, and further refinements were made through team discussion. Following this process, the Traditional Chinese version of the 15-STARS was finalized for field testing in community pharmacies. Field-Testing Phase Study Sites Six community pharmacies collaborating with National Taiwan University were selected as study sites. These pharmacies were located in northern Taiwan (i.e, Taipei City, New Taipei City, and Hsinchu City) and central Taiwan (i.e, Taichung City) to enhance demographic diversity among participants. Participant Recruitment Sample size considerations were informed by recommendations for psychometric evaluation, which specify a minimum of 10 participants per item.26 Accordingly, a minimum sample size of 150 participants was targeted for the 15-item instrument. Participants were recruited using convenience sampling. Eligibility criteria included being 18 years of age or older, having a diagnosis of T2D, being prescribed at least one glucose-lowering medication for daily use, and being able to read and communicate in Mandarin. The National Health Insurance MediCloud System was used to verify glucose-lowering medications and retrieve the most recent A1C (HbA1c) values. Exclusion criteria included inability to understand Mandarin, unavailability of medication or clinical information through the MediCloud System, or cognitive impairment that could hinder questionnaire completion. At each site, pharmacists screened patients presenting for prescription refills and referred eligible patients to a trained researcher. The researcher accompanied participants to a private area, provided a study information sheet, and allowed 5–10 minutes for review. Written informed consent was obtained from all participants, who received a copy of the consent form for their records. The questionnaire was then administered face-to-face and took approximately 5–10 minutes to complete. Questionnaire administration could occur either before or after prescription dispensing, depending on participant preference. Upon completion, participants received a pharmacy coupon valued at NT$200 (approximately US$7) as compensation. Measures and Data Collection In addition to the 15-STARS items, demographic and clinical variables were collected. The Traditional Chinese version of the ChARMS-T was included to assess criterion validity.27 The full questionnaire consisted of 33 items and was administered face-to-face. Sociodemographic and Clinical Characteristics Participants were asked to self-report sociodemographic information, including year of birth, sex, educational attainment, and annual household income. Clinical characteristics, such as the number of medications prescribed for diabetes, refill dates within the previous 90 days, and the most recent A1C value, were extracted from the MediCloud System. The medication regimen complexity index (MRCI) and proportion of days covered (PDC) were calculated using these data.28 An A1C level of 7.0% or lower is generally recommended for many adults with diabetes, although individualized glycemic targets may be appropriate.29 The 15-Item Screening Tool for Adherence to Medicines The 15-STARS is a non-disease-specific, patient-reported questionnaire consisting of four conceptual sections. These include 5 items assessing barriers to medication handling, 6 items addressing reasons for non-adherence, 3 items evaluating medication discontinuation behaviors, and 1 item exploring the need for pharmacist support.21 Except for the item assessing the frequency of missed doses (Item 13), all items use a forced-choice yes-no response format to minimize response burden. The instrument was developed primarily as a multidimensional screening tool in which individual items identify specific medication-taking problems and potential adherence barriers, instead of functioning as a unidimensional scale intended to generate a single overall adherence score.21 The Traditional Chinese Version of the Adherence to Refills and Medications Scale The ChARMS-T includes 12 items that evaluate medication-taking behavior (8 items) and refill adherence (4 items). Each item is rated on a 4-point Likert scale ranging from 1 (none of the time) to 4 (all of the time), yielding total scores ranging from 12 to 48, with lower scores reflecting better adherence. The scale has shown acceptable internal consistency, with McDonald’s omega (ω) values of 0.841 for medication-taking and 0.647 for medication refill, as well as evidence of criterion validity among individuals with T2D.8 Data Analysis Participant characteristics were summarized using descriptive statistics. Continuous variables were reported as means and standard deviations (SDs), while categorical variables were presented as frequencies and percentages. Content validity of the translated 15-STARS was assessed using the content validity index (CVI), with an item-level CVI (I-CVI) greater than 0.78 and a scale-level CVI (S-CVI) greater than 0.80 considered acceptable.30 Criterion validity was examined at the item level by correlating selected 15-STARS items with conceptually corresponding ChARMS-T items or relevant clinical measures. This approach was used because the 15-STARS and ChARMS-T assess different dimensions of medication use and do not provide directly equivalent total scores.8,21 Items were selected only when a conceptually comparable criterion was available, including forgetfulness (Item 10), discontinuation of medication when feeling better (Item 12), and frequency of missed medication doses (Item 13). Other 15-STARS items were not included in the criterion validity analysis because no sufficiently comparable item or external criterion was available. Internal consistency was evaluated using McDonald’s ω for Items 6–9, which assess conceptually related belief-based reasons for non-adherence, including concerns about side effects, taking too many medications, medication dependence, and perceived lack of benefit. Internal consistency was not calculated for the entire 15-STARS because the instrument was designed as a multidimensional screening tool comprising items that assess distinct medication-taking behaviors, barriers, and support needs, rather than as a measure of a single latent construct. A McDonald’s ω value of 0.6 or higher was considered acceptable.31 Descriptive statistics, criterion validity coefficients, and McDonald’s ω were computed using SPSS version 29. All statistical tests were two-sided, and a p-value of less than 0.05 was considered statistically significant. Results Participants’ Characteristics During the cognitive debriefing phase, ten pharmacists participated in the evaluation, with equal numbers of men and women and a mean age of 36.6 years (SD = 6.98). Five pharmacists held a bachelor’s degree, while the other five held a master’s degree (Table 1). In addition, 10 patients (5 men and 5 women) participated in the cognitive debriefing, with a mean age of 55.4 years (SD = 19.44). Most patients (n = 6, 60.0%) had attained at least a high school education (Table 1).Table 1Demographic and Clinical Characteristics of Participants in the Cognitive Debriefing StageVariablesPharmacists (n =10)Patients (n =10)n (%)Mean (SD)n (%)Mean (SD)Age (years)36.6 (6.98)55.4 (19.44)Gender Female5 (50.0)5 (50.0) Male5 (50.0)5 (50.0)Education Elementary school or below0 (0)1 (10.0) Junior high school0 (0)3 (30.0) High school0 (0)4 (40.0) Junior college0 (0)0 (0) Bachelor’s degree or a 4-year college degree5 (50.0)1 (10.0) Master’s degree or above5 (50.0)1 (10.0) Table 1Demographic and Clinical Characteristics of Participants in the Cognitive Debriefing StageVariablesPharmacists (n =10)Patients (n =10)n (%)Mean (SD)n (%)Mean (SD)Age (years)36.6 (6.98)55.4 (19.44)Gender Female5 (50.0)5 (50.0) Male5 (50.0)5 (50.0)Education Elementary school or below0 (0)1 (10.0) Junior high school0 (0)3 (30.0) High school0 (0)4 (40.0) Junior college0 (0)0 (0) Bachelor’s degree or a 4-year college degree5 (50.0)1 (10.0) Master’s degree or above5 (50.0)1 (10.0) Demographic and Clinical Characteristics of Participants in the Cognitive Debriefing Stage In the field-testing phase, 238 patients were approached, and 215 completed the study, resulting in a response rate of 90.3%. Participants ranged in age from 34 to 88 years, with a mean age of 65.28 years (SD = 9.99), and 50.7% (n = 109) were women. Fewer than one quarter of participants (n = 51, 23.7%) had a bachelor’s degree or higher, whereas more than half (n = 118, 54.9%) reported an annual household income above NT$660,000. The mean A1C level was 7.03% (SD = 1.11; range: 5.0–14.9), and the most recent A1C measurement had been obtained a mean of 54.10 days before survey completion (SD = 34.94; range: 1–165 days) (Table 2).Table 2Demographic and Clinical Characteristics of Participants in the Field-Testing Phase (n = 215)Variablesn (%)Mean (SD)Age65.28 (9.99)Gender Female109 (50.7) Male106 (49.3)Education Elementary school or below35 (16.3) Junior high school24 (11.2) High school62 (28.8) Junior college43 (20.0) Bachelor’s degree or a 4-year college degree39 (18.1) Master’s degree or above12 (5.6)Annual household income (NTD) Less than 350,00054 (25.1) 350,001–659,99943 (20.0) 660,000–959,99953 (24.7) 960,000–1,309,99939 (18.1) 1,310,000–2,210,00019 (8.8) Above 2,210,0007 (3.3)A1C (%)7.03 (1.11) Less than or equal to 7% (53.0 mmol/mol)140 (65.1) More than 7% (53.0 mmol/mol)75 (34.9)Proportion of days covered (%)92.57 (6.21) Missing3 Table 2Demographic and Clinical Characteristics of Participants in the Field-Testing Phase (n = 215)Variablesn (%)Mean (SD)Age65.28 (9.99)Gender Female109 (50.7) Male106 (49.3)Education Elementary school or below35 (16.3) Junior high school24 (11.2) High school62 (28.8) Junior college43 (20.0) Bachelor’s degree or a 4-year college degree39 (18.1) Master’s degree or above12 (5.6)Annual household income (NTD) Less than 350,00054 (25.1) 350,001–659,99943 (20.0) 660,000–959,99953 (24.7) 960,000–1,309,99939 (18.1) 1,310,000–2,210,00019 (8.8) Above 2,210,0007 (3.3)A1C (%)7.03 (1.11) Less than or equal to 7% (53.0 mmol/mol)140 (65.1) More than 7% (53.0 mmol/mol)75 (34.9)Proportion of days covered (%)92.57 (6.21) Missing3 Demographic and Clinical Characteristics of Participants in the Field-Testing Phase (n = 215) On average, participants were prescribed 2.00 (SD = 1.00) glucose-lowering medications, and 71.2% (n = 153) had a dosing frequency of at least twice daily. Two participants (0.9%) used injectable medications only, while 14 (6.5%) were prescribed both oral and injectable medications. Metformin and insulin were the most commonly prescribed oral and injectable glucose-lowering medications, respectively (Table 3). The mean PDC was 92.57% (SD = 6.21) (Table 2).Table 3Characteristics of Participants’ Diabetes Management MedicationsVariablesn (%)Mean (SD)Number of diabetes medications2.00 (1.00)Medication regimen complexity index6.12 (3.62)Dosing frequency of diabetes medications Once daily61 (28.4) Twice daily123 (57.2) Three times daily29 (13.5) Four times daily1 (0.5) Once every four weeks1 (0.5)Oral medications α-glucosidase inhibitors9 (4.2) Biguanides136 (63.3) Thiazolidinediones25 (11.6) Sulfonylurea54 (25.1) Meglitinides6 (2.8) Sodium-glucose co-transporter 2 inhibitors41 (19.0) Dipeptidyl peptidase 4 inhibitors46 (21.4) Combination drugs66 (30.7) DPP4 inhibitors + Biguanides24 (11.2) Thiazolidinediones + Biguanides6 (2.8) Sulfonylurea + Biguanides20 (9.3) SGLT2 inhibitors + Biguanides8 (3.7) SGLT2 inhibitors + DPP4 inhibitors37 (17.2) DPP4 inhibitors + Thiazolidinediones1 (0.5)Injectable medications16 (7.4) Insulin12 (5.6) Rapid-acting2 (9.3) Insulin aspart1 (0.5) Insulin lispro1 (0.5) Long-acting6 (2.8) Insulin glargine5 (2.3) Insulin degludec1 (0.5) Insulin degludec + aspart4 (1.9) Glucagon-like peptide 1 receptor agonists3 (1.4) GLP1 agonist + insulin1 (0.5) Table 3Characteristics of Participants’ Diabetes Management MedicationsVariablesn (%)Mean (SD)Number of diabetes medications2.00 (1.00)Medication regimen complexity index6.12 (3.62)Dosing frequency of diabetes medications Once daily61 (28.4) Twice daily123 (57.2) Three times daily29 (13.5) Four times daily1 (0.5) Once every four weeks1 (0.5)Oral medications α-glucosidase inhibitors9 (4.2) Biguanides136 (63.3) Thiazolidinediones25 (11.6) Sulfonylurea54 (25.1) Meglitinides6 (2.8) Sodium-glucose co-transporter 2 inhibitors41 (19.0) Dipeptidyl peptidase 4 inhibitors46 (21.4) Combination drugs66 (30.7) DPP4 inhibitors + Biguanides24 (11.2) Thiazolidinediones + Biguanides6 (2.8) Sulfonylurea + Biguanides20 (9.3) SGLT2 inhibitors + Biguanides8 (3.7) SGLT2 inhibitors + DPP4 inhibitors37 (17.2) DPP4 inhibitors + Thiazolidinediones1 (0.5)Injectable medications16 (7.4) Insulin12 (5.6) Rapid-acting2 (9.3) Insulin aspart1 (0.5) Insulin lispro1 (0.5) Long-acting6 (2.8) Insulin glargine5 (2.3) Insulin degludec1 (0.5) Insulin degludec + aspart4 (1.9) Glucagon-like peptide 1 receptor agonists3 (1.4) GLP1 agonist + insulin1 (0.5) Characteristics of Participants’ Diabetes Management Medications Validity and Reliability In the first round, I-CVI values ranged from 0.40 to 1.00 across items and their accompanying text. The mean S-CVI values were 0.91 for comprehensiveness, 0.93 for linguistic style, 0.95 for clarity, and 0.98 for ease of understanding. Based on feedback from the expert panel, all items were subsequently revised. In the second round, both I-CVI and S-CVI values reached 1.00, indicating that all items met the prespecified criteria for content validity. These findings indicated that the 15-STARS demonstrated satisfactory content validity. Regarding criterion validity, Item 10 of the 15-STARS, which assessed forgetfulness as a barrier to medication adherence, showed a moderate correlation with the corresponding item of the ChARMS-T (r = 0.505). Similarly, Item 12, addressing discontinuation of medication when feeling well, was correlated with the corresponding item of the ChARMS-T (r = 0.36). Item 13, assessing the frequency of missed medication doses, showed a strong correlation with the ChARMS-T medication-taking subscale (r = 0.635) but was not significantly associated with PDC. McDonald’s ω was 0.831 for Items 6–9 of the 15-STARS, indicating good internal consistency among these conceptually related items. Adherence Level and Potential Barriers Among the 215 participants, 114 (53.0%) reported not missing any medication doses during the preceding two weeks. Sixty-four participants (29.8%) reported missing one to two doses, 31 (14.4%) reported missing three to four doses, two (0.9%) reported missing five or more doses, and four (1.9%) reported missing medication daily. In the preceding six months, 36 participants (16.7%) reported discontinuing at least one medication without informing their physician, and 40 (18.6%) indicated that they would stop taking medication when they felt well. Twenty-five participants (11.6%) reported at least one difficulty with medication use, although fewer than 4% reported any specific individual problem. Nearly half of the participants (n = 102, 47.4%) reported at least one reason for non-adherence. The most frequently reported reasons were forgetfulness (n = 71, 33.0%), the number of prescribed medications (n = 40, 18.6%), and concerns about potential side effects (n = 33, 15.3%) (Table 4). In addition, 103 participants (47.9%) expressed interest in receiving additional support from pharmacists.Table 4Responses to the Traditional Chinese Version of the 15-STARS Questionnaire (n = 215)ItemsYes (%)No (%)Barriers to medication handling Q1. I recognise all the medicines I use (by their shapes, their colours, their sizes etc)208 (96.7)7 (3.3) Q2. I have difficulties getting my medicines ready before I use them (e.g, popping them out of a packaging, halving tablets, counting drops)8 (3.7)207 (96.3) Q3. I understand the instructions for using for all the medicines I use210 (97.7)5 (2.3) Q4. I have problems with the use of at least one of my medicines (e.g, swallowing pills, using the inhaler device, injection)8 (3.7)207 (96.3) Q5. I get all my medicines before I run out of them208 (96.7)7 (3.3)Reasons for non-adherence Q6. Because I am concerned about possible side effects of some of my medicines33 (15.3)182 (84.7) Q7. Because I am concerned that I am taking too many medicines40 (18.6)175 (81.4) Q8. Because I am concerned about becoming dependent on some of my medicines28 (13.0)187 (87.0) Q9. Because I believe I have no benefits from some of my medicines20 (9.3)195 (90.7) Q10. Because I forget to take some of my medicines71 (33.0)144 (67.0) Q11. Because of the price of some of my medicines6 (2.8)209 (97.2)Medication discontinuation behaviors Q12. Do you stop taking some of your medicines if you are feeling well?40 (18.6)175 (81.4) Q13. Over the past 2 weeks, how often have you missed taking any of your medicines? 0 times: 114 (53.0%); 1–2 times: 64 (29.8%); 3–4 times: 31 (14.4%); 5 or more times: 2 (0.9%); everyday: 4 (1.9%) Q14. Over the past 6 months, have you ever stopped taking any of your medicines without informing your doctor?36 (16.7)179 (83.3)The need for pharmacist support Q15. Would you like some help from the pharmacist?103 (47.9)112 (52.1) Table 4Responses to the Traditional Chinese Version of the 15-STARS Questionnaire (n = 215)ItemsYes (%)No (%)Barriers to medication handling Q1. I recognise all the medicines I use (by their shapes, their colours, their sizes etc)208 (96.7)7 (3.3) Q2. I have difficulties getting my medicines ready before I use them (e.g, popping them out of a packaging, halving tablets, counting drops)8 (3.7)207 (96.3) Q3. I understand the instructions for using for all the medicines I use210 (97.7)5 (2.3) Q4. I have problems with the use of at least one of my medicines (e.g, swallowing pills, using the inhaler device, injection)8 (3.7)207 (96.3) Q5. I get all my medicines before I run out of them208 (96.7)7 (3.3)Reasons for non-adherence Q6. Because I am concerned about possible side effects of some of my medicines33 (15.3)182 (84.7) Q7. Because I am concerned that I am taking too many medicines40 (18.6)175 (81.4) Q8. Because I am concerned about becoming dependent on some of my medicines28 (13.0)187 (87.0) Q9. Because I believe I have no benefits from some of my medicines20 (9.3)195 (90.7) Q10. Because I forget to take some of my medicines71 (33.0)144 (67.0) Q11. Because of the price of some of my medicines6 (2.8)209 (97.2)Medication discontinuation behaviors Q12. Do you stop taking some of your medicines if you are feeling well?40 (18.6)175 (81.4) Q13. Over the past 2 weeks, how often have you missed taking any of your medicines? 0 times: 114 (53.0%); 1–2 times: 64 (29.8%); 3–4 times: 31 (14.4%); 5 or more times: 2 (0.9%); everyday: 4 (1.9%) Q14. Over the past 6 months, have you ever stopped taking any of your medicines without informing your doctor?36 (16.7)179 (83.3)The need for pharmacist support Q15. Would you like some help from the pharmacist?103 (47.9)112 (52.1) Responses to the Traditional Chinese Version of the 15-STARS Questionnaire (n = 215) Bivariate correlation analyses indicated that greater regimen complexity (rs = 0.144, p = 0.035) and higher dosing frequency (rs = 0.168, p = 0.014) were significantly associated with lower adherence. Poorer glycemic control was significantly associated with greater regimen complexity (rs = 0.415, p < 0.001), a higher number of prescribed glucose-lowering medications (rs = 0.378, p < 0.001), more frequent dosing (rs = 0.246, p < 0.001), and lower adherence (rs = 0.224, p < 0.001) (Table 5).Table 5Correlation Matrix of Diabetes Medication Adherence, A1C, and Other Variables(1)(2)(3)(4)(5)(6)(7)(8)(9)Age (1)1.000Gender (2)a−0.0841.000Education (3)−0.283***0.322***1.000Annual household income (4)−0.158*0.160*0.456***1.000Medication Regimen Complexity (5)−0.0660.0920.0100.0051.000Number of diabetes medications (6)−0.0100.0910.0280.0050.883***1.000Frequency of diabetes medications (7)−0.0710.085−0.0050.0740.570***0.370***1.000Item 13: frequency of missed medication doses (8)−0.134−0.0840.005−0.0200.144*0.1240.168*1.000A1C level (9)−0.074−0.0380.0750.0160.415***0.378***0.246**0.224***1.000Notes: aFemale was used as the comparison group, * p < 0.05, ** p < 0.01, *** p < 0.001. Table 5Correlation Matrix of Diabetes Medication Adherence, A1C, and Other Variables(1)(2)(3)(4)(5)(6)(7)(8)(9)Age (1)1.000Gender (2)a−0.0841.000Education (3)−0.283***0.322***1.000Annual household income (4)−0.158*0.160*0.456***1.000Medication Regimen Complexity (5)−0.0660.0920.0100.0051.000Number of diabetes medications (6)−0.0100.0910.0280.0050.883***1.000Frequency of diabetes medications (7)−0.0710.085−0.0050.0740.570***0.370***1.000Item 13: frequency of missed medication doses (8)−0.134−0.0840.005−0.0200.144*0.1240.168*1.000A1C level (9)−0.074−0.0380.0750.0160.415***0.378***0.246**0.224***1.000Notes: aFemale was used as the comparison group, * p < 0.05, ** p < 0.01, *** p < 0.001. Correlation Matrix of Diabetes Medication Adherence, A1C, and Other Variables Notes: aFemale was used as the comparison group, * p < 0.05, ** p < 0.01, *** p < 0.001. Discussion This study provides initial evidence supporting selected measurement properties and the practical utility of the Traditional Chinese version of the 15-STARS in Taiwanese community pharmacy settings. Content validity was satisfactory, selected items were associated with corresponding self-reported adherence measures, and Items 6–9 demonstrated good internal consistency. These findings support the potential use of the 15-STARS as a practical screening tool for identifying patients at risk of nonadherence and exploring adherence-related barriers during routine pharmacy encounters. Item 13 was not significantly associated with PDC, which may reflect differences in the aspects and timeframes of adherence captured by the two measures. Proportion of days covered reflects medication availability based on prescription refills over a 3-month period, whereas Item 13 assesses missed doses during the preceding 2 weeks. This finding is consistent with the original developers’ description of the 15-STARS as capturing dimensions of medication adherence that differ from those assessed using conventional refill-based measures.21 The high completion rate and brief administration time observed in the field-testing phase suggest that the translated 15-STARS could be incorporated into routine pharmacy encounters with minimal disruption to workflow. Both pharmacists and patients found the items clear and relevant, further supporting the instrument’s usability in real-world practice. The present study complements our previously published adherence-related research. While earlier studies focused on specific constructs such as medication-taking behavior, medication beliefs, or medication understanding,8,22,23 the 15-STARS integrates multiple adherence-related domains with a focus on modifiable derterminants into a single screening instrument. This broader scope may facilitate the identification of actionable barriers during routine pharmacist-patient interactions and support the development of individualized adherence interventions. From an implementation perspective, the translated 15-STARS offers added value by simultaneously identifying medication-taking problems and specific, modifiable barriers to adherence.21 In this study, nearly half of participants reported at least one reason for non-adherence, with forgetfulness, regimen complexity, and concerns about side effects emerging as the most common challenges. These barriers align closely with intervention strategies already within pharmacists’ scope of practice, such as medication synchronization, discussions with prescribers about regimen simplification, reminder strategies, and targeted counseling on adverse effects.32 The structured nature of the 15-STARS allows pharmacists to quickly prioritize issues during brief consultations and tailor follow-up support to individual patient needs, providing more information than general adherence inquiries.21 The lack of a significant association between Item 13 and PDC warrants careful consideration when interpreting the validity of the 15-STARS as a measure of medication adherence. Although selected 15-STARS items were associated with corresponding ChARMS-T measures, convergence with the refill-based adherence measure was not demonstrated. Refill data provide system-level signals but may fail to capture intentional non-adherence, beliefs about medicines, or day-to-day challenges that influence behavior.33 Conversely, self-reported measures may be affected by recall and social desirability bias. Thus, self-reported and refill-based measures capture different aspects of medication use and may provide complementary information.34 The value of the 15-STARS may therefore lie primarily in identifying patients at risk of nonadherence and eliciting the practical, behavioral, and perceptual barriers underlying their medication-taking behavior. In practice, integrating the 15-STARS with existing electronic health information systems, such as MediCloud, could allow pharmacists to combine patient-reported information with refill-based measures to obtain a more comprehensive understanding of medication use and identify patients who may benefit from additional support. Contextual factors identified in this study further inform potential implementation strategies. Although the prevalence of non-adherence (47.0%) was comparable to previous studies,35 cost-related barriers were uncommon in the present study, which may partly reflect Taiwan’s universal health insurance system.36 In contrast, belief-related concerns and forgetfulness predominated, suggesting that adherence interventions in this setting may need to prioritize behavioral and educational components.37 Behavioral and educational strategies, including reminder systems, regimen simplification, and structured pharmacist consultations, may help address these challenges.32,38,39 Cultural perceptions of Western medications, including concerns about long-term safety and side effects, may influence medication-taking decisions and should be addressed during pharmacist-patient interactions.40,41 These findings support adapting adherence interventions to the local context instead of directly transferring approaches developed in other healthcare systems. The high proportion of participants expressing interest in additional pharmacist support suggests receptiveness to expanded adherence services in community pharmacies.42 Support from healthcare professionals has been recognized as an important motivator for adherence, and understanding the underlying reasons for nonadherence is essential for developing effective interventions. Beyond screening, the 15-STARS could serve as an operational prompt for pharmacists to initiate structured follow-up actions, including brief counseling, motivational interviewing, or referral to collaborative care programs.43,44 Given its lack of licensing fees and brief administration time, the instrument may be suitable for broader implementation in Taiwan’s community pharmacy system, where efficiency and cost-effectiveness are critical considerations. Future implementation studies should evaluate how the translated 15-STARS performs when embedded in routine chronic disease management services, including its impact on workflow efficiency, intervention fidelity, patient engagement, and longer-term clinical outcomes. Strengths and Limitations A key strength of this study lies in its emphasis on real-world applicability, with field testing conducted in active community pharmacies and data collected during routine patient visits. The inclusion of both pharmacists and patients throughout the translation and cross-cultural adaptation process enhanced the practical relevance, clarity, and acceptability of the instrument. The Traditional Chinese 15-STARS demonstrated satisfactory content validity, selected item-level associations with another self-reported adherence measure, and good internal consistency among conceptually related Items 6–9, providing initial evidence for selected measurement properties of the translated instrument. Its simple response format minimizes respondent burden while capturing information on medication-taking behavior, patient-reported experiences, and adherence-related barriers. By facilitating structured communication between healthcare providers and patients, the 15-STARS may support the identification of adherence barriers and inform the delivery of tailored, patient-centered interventions. In addition to its clinical utility, the instrument may serve as a useful research tool for evaluating adherence-focused services implemented in routine practice. Several limitations relevant to implementation and psychometric evaluation should be considered when interpreting the findings. Participants were recruited using convenience sampling from community pharmacies located primarily in metropolitan areas and partnered with a single academic institution. This recruitment approach may have favored individuals who were more engaged with pharmacy services and may limit generalizability to other practice settings, such as hospitals or rural pharmacies, where workflows and adherence challenges may differ. The cross-sectional design precluded evaluation of longitudinal measurement and implementation outcomes, including sustained use, sensitivity to change, and the impact of repeated administration on adherence over time. In addition, test-retest reliability could not be assessed due to recruitment constraints and time limitations, limiting conclusions regarding the stability of responses over time. The psychometric findings should also be interpreted in the context of the measurement structure of the 15-STARS. Internal consistency was evaluated only for Items 6–9 because these items assess conceptually related medication-related beliefs as reasons for nonadherence, whereas the remaining items assess distinct practical, behavioral, and support-related aspects of medication use. Thus, the McDonald’s ω of 0.831 should not be interpreted as evidence of internal consistency for the entire 15-STARS. Similarly, criterion validity was evaluated only for selected items for which conceptually corresponding ChARMS-T measures were available. Although these items showed weak-to-moderately strong correlations with corresponding self-reported adherence measures, Item 13 was not significantly associated with PDC, providing limited evidence of convergence with the refill-based adherence measure. Because the 15-STARS comprises heterogeneous screening items and was not designed to measure a single latent construct, structural validity using exploratory or confirmatory factor analysis was not evaluated in the present study. The present study also did not evaluate the diagnostic accuracy of the Traditional Chinese 15-STARS for classifying patients as adherent or nonadherent. An independent reference standard suitable for evaluating diagnostic accuracy was not incorporated into the study design; therefore, sensitivity, specificity, positive and negative predictive values, receiver operating characteristic analysis, area under the curve, and an optimal cut-off value were not determined. Accordingly, the present findings should not be interpreted as establishing the diagnostic performance of the Traditional Chinese 15-STARS for identifying medication nonadherence. Instead, they provide initial support for its potential utility in identifying self-reported medication-taking problems and adherence barriers. Future studies using an appropriate external reference standard are needed to further evaluate its screening performance. As with all self-reported measures, responses may be influenced by recall and social desirability bias, potentially affecting accuracy. In addition, non-response bias cannot be ruled out, as not all approached patients agreed to participate. Future research should evaluate additional measurement properties, screening performance, implementation fidelity, pharmacist training requirements, and the effectiveness of 15-STARS-guided interventions on patient-reported and clinical outcomes across diverse community pharmacy and healthcare settings. Conclusions The Traditional Chinese 15-STARS demonstrated satisfactory content validity, evidence of criterion validity for selected items, and internal consistency among conceptually related belief-based items. The findings support its potential use as a practical screening tool for identifying patients at risk of medication nonadherence and exploring adherence-related barriers in community pharmacy practice, but not as a definitive measure of actual medication adherence. Its use in clinical practice may help healthcare professionals identify patients who could benefit from tailored education and individualized support to promote adherence. This study also identified key adherence challenges among Taiwanese patients with T2D, including forgetfulness, concerns about taking too many medications, and concerns about medication side effects. These findings provide a foundation for developing patient-centered adherence support approaches. Further validation against appropriate objective or external reference measures is needed to establish the screening performance of the Traditional Chinese 15-STARS and its applicability across different diseases, patient populations, and healthcare settings.