Development and reliability testing of BRANT-MERQS-2A: A quality assessment tool for medication review type 2a reports Authors: Eline Tobback, Annouk Peeters, Nabileh Basir, Emilie Van Haver, Juno Quayhackx, Maja Brumer, Ronja Woltersdorf, Guido R. Y. De Meyer, Hans De Loof Published: 2026 ABSTRACT IntroductionMedication reviews (MRs) are increasingly implemented to optimize medication use and address drug-related problems in patients with complex pharmacotherapy, with medication review type 2a (MR2a) representing an intermediate format involving a pharmacist-led structured assessment with patient interview, based on clinical and medication data, without direct prescriber interaction. Despite their widespread implementation, systematic and operational approaches to evaluate the quality of MRs remain limited. In Belgium, Medication Review type 2a (MR2a) has been implemented in routine practice of community pharmacies since 2023, yet no validated instruments exist to evaluate the quality of written MR2a reports.AimThis study aimed to adapt the existing BRANT-MERQS framework for the quality evaluation of MR2a reports and to assess the usability, reproducibility and reliability of the resulting tool (BRANT-MERQS-2A).MethodRelevant criteria were selected from the original BRANT-MERQS and expanded with criterion-specific instructions to form the BRANT-MERQS-2A. Initial testing involved two student assessors independently evaluating 98 MR2a reports, with interrater agreement (IRA) assessed using weighted Cohen’s Kappa. In a subsequent extended testing phase, the same reports were assessed by 61 final-year pharmacy students, with three randomly selected students evaluating each MR2A report. Interrater reliability was evaluated using the Intraclass Correlation Coefficient (ICC) and Gwet’s AC1. Assessment duration was recorded to evaluate feasibility.ResultsInitial testing demonstrated strong IRA between the two student assessors. In the extended student cohort, agreement at the criterion level ranged from poor to moderate, with an overall ICC of 0.62, indicating moderate reliability. Criteria differed in discriminatory capacity, with some showing high absolute agreement but limited variability. Assessment time decreased over successive rounds, reflecting a learning effect and improved efficiency.ConclusionThe BRANT-MERQS-2A provides a feasible and structured approach for assessing the quality of MR2a reports. While variability at criterion level reflects the inherent complexity of MR2a, the tool demonstrates potential for self-and peer-assessment and offers a valuable foundation for further refinement and integration into routine quality monitoring.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s11096-026-02174-4. Impact statements BRANT-MERQS-2A provides a structured framework to assess quality of MR2a reports, supporting more consistent documentation and evaluation of medication review services in community pharmacy.Use of BRANT-MERQS-2A may support pharmacists in identifying areas for improvement in medication review practice and stimulate continuous quality improvement in pharmaceutical care.Improving the quality of medication reviews may enhance the identification and management of drug-related problems, contributing to safer and more effective medication use for patients. BRANT-MERQS-2A provides a structured framework to assess quality of MR2a reports, supporting more consistent documentation and evaluation of medication review services in community pharmacy. BRANT-MERQS-2A provides a structured framework to assess quality of MR2a reports, supporting more consistent documentation and evaluation of medication review services in community pharmacy. Use of BRANT-MERQS-2A may support pharmacists in identifying areas for improvement in medication review practice and stimulate continuous quality improvement in pharmaceutical care. Use of BRANT-MERQS-2A may support pharmacists in identifying areas for improvement in medication review practice and stimulate continuous quality improvement in pharmaceutical care. Improving the quality of medication reviews may enhance the identification and management of drug-related problems, contributing to safer and more effective medication use for patients. Improving the quality of medication reviews may enhance the identification and management of drug-related problems, contributing to safer and more effective medication use for patients. Introduction Over the past decades, pharmacists have evolved from primarily dispensing and compounding medications toward providing clinical advice, patient education and comprehensive pharmaceutical care [1, 2]. According to the Pharmaceutical Care Network Europe (PCNE), pharmaceutical care is defined as the pharmacist’s contribution to the care of individuals, in order to optimize medicines use and improve health outcomes [3]. This shift is particularly relevant in the context of an aging population, where increasing life expectancy is accompanied by a higher prevalence of chronic diseases and, consequently, more complex medication regimens and polypharmacy [4–8]. To address the challenges with these complex regimens, medication review (MR) has become an increasingly adopted strategy [9–11]. The PCNE defines MR as a structured evaluation of a patient’s medications to optimize medication use and improve outcomes by identifying and addressing drug-related problems (DRPs) [12, 13]. The PCNE further classifies MR into a simple form (Type 1), and intermediate form (Type 2) and a comprehensive form (Type 3) [12, 14]. In multiple countries, different types of MR have been integrated into routine practice, although the extent of their adoption varies considerably [15–18]. Multiple studies demonstrate that MR interventions are effective in identifying, preventing and resolving DRPs, highlighting their potential value in improving medication management [18–20]. However, evidence supporting their impact on patient outcomes remains inconsistent, meaning that the overall clinical benefit and effectiveness of this service cannot yet be fully determined [9, 18, 19, 21, 22]. In addition, uncertainty persists regarding the quality of MR delivery, primarily due to limited reporting on essential intervention components and fidelity, hindering interpretation of their effectiveness [18]. Medication Review Type 2a (MR2a) has been implemented and reimbursed in Belgium since April 1, 2023 [23]. To support transparency and standardized documentation of this service, FarmaFlux developed an online electronic documentation form (e-form), integrated into community pharmacy software [24, 25]. FarmaFlux is a non-profit umbrella organization responsible for secure and uniform exchange of data between pharmacies and supervising the quality and privacy of the information stored in the Shared Pharmaceutical Record (SPR) [25, 26]. The e-form guides pharmacists through the MR2a process while systematically capturing key elements of the intervention [24]. It automatically loads predefined patient and medication data, allowing pharmacists to complete additional fields and to specify which medications are considered during the MR2a [24]. Despite the availability of this tool, MR2a uptake has remained relatively stable and occasional concerns regarding service quality persist [27]. Assessing the quality of MR is challenging due to the multifaceted nature of the intervention and the absence of a universally accepted definition of quality [28–30]. Systematic monitoring of care processes is essential for sustainability, as quality measurement and standardized indicators enhance transparency, identify variation in practice, and support targeted quality improvement [31]. Quality controls are therefore particularly relevant for MR implementation, as they contribute to patient safety, build trust among stakeholders and support long-term service sustainability [32–39]. Although international organizations have proposed conceptual quality frameworks for community pharmacy, their practical applicability remains limited due to their theoretical nature and the lack of operational tools to monitor clinical services, such as MR [40, 41]. Despite extensive research demonstrating the potential clinical and economic value of MR, insight into professional standardization and real-time quality assurance remains limited [27, 42, 43]. Recent studies highlight significant variability in MR delivery, underscoring the lack of structured frameworks to monitor process quality [27, 42]. While PCNE indicators provide a useful benchmark for process and structural aspect of services, such as MR, they do not capture clinical quality at the individual patient level and cannot determine whether the most relevant pharmacotherapeutic problems are identified or resolved, meaning high process scores do not guarantee optimal clinical outcomes [40, 44]. In a previous study, a quality assessment tool was developed specifically for MR3, the BRussels ANTwerp Medication Review Quality Score (BRANT-MERQS), which evaluates MR quality based on a set of predefined criteria [45, 46]. Although MR involves additional elements beyond the written report, this study focuses solely on this key part of the procedure [45]. Variability between assessors observed in the original study highlighted the need for a detailed scoring manual with clear guidance [45]. However, MR3 has not yet been systematically implemented in Belgium, in contrast to MR2a. To date, there are no comparable tools available for this type of MR. Given the growing emphasis on quality in healthcare, the development of robust quality measures for MR services is essential to ensure consistent, high-level delivery and to support the sustainable implementation and reimbursement of these services [33, 35–39]. Aim This study aims to adapt the existing BRANT-MERQS approach for MR2a quality evaluation and assess its usability, reproducibility and reliability. Method Development of the modified BRANT-MERQS and the scoring manual The Brussels-Antwerp Medication Review Quality Score for MR2a (BRANT-MERQS-2A) was modified from the original BRANT-MERQS [45]. Criteria potentially applicable to MR2a were independently screened for relevance by four researchers (ET, YQ, NB and AP), resulting in a preliminary list. Following consensus meetings including a fifth researcher (HDL), eleven criteria were selected based on their applicability to the Belgian MR2a context and their compatibility to the current e-form. Selection was conducted through iterative discussions, whereby discrepancies were resolved through deliberation until full consensus was reached for each criterion. To minimize inter-rater variability, a scoring manual with explicit, criterion-specific instructions for each level of the 0–3 scale (0: absence; 3 high quality) was developed. Consensus on wording and scoring structure was reached among ET, HDL, AP and NB, resulting in the initial BRANT-MERQS-2A version. Data collection The MR2a reports analyzed were anonymized PDF outputs automatically generated from the e-form, used to document MR2a performed by pharmacy students during their final internship. Final-year students are required to conduct at least four MR2a with their internship mentor. These reports were submitted as part of a mandatory internship assignment, and students could consent to their use for research. The MR2a’s were not restricted to a specific patient population, disease group, or demographic characteristics, but were conducted in accordance with the Belgian reimbursement criteria for MR2a, including, among others the use of at least five chronically reimbursed medications [23]. As such, the sample reflects routine practice rather than a predefined clinical subgroup. This data collection was routinely conducted as part of students’ final internship assignments. The initial testing phase took place between March 2025 and June 2025. Following the decision to pursue publication, an ethics committee application was initiated in October 2025 as a precautionary measure for the final testing phase, in which students assessed MR2a reports authored by peers. Due to the illness of ET, final approval from the ethics committee was obtained in January 2026. The final testing phase and subsequent analysis of the data were conducted between January and March 2026. Pilot testing and initial testing of BRANT-MERQS-2A A pilot phase was conducted to evaluate the clarity and preliminary usability of the BRANT-MERQS-2A scoring instrument prior to full application. An initial sample of 20 MR2a reports was independently assessed by two student assessors (NB and AP) using the preliminary version of BRANT-MERQS-2A. Following independent scoring, discrepancies between assessors were reviewed and discussed until consensus was achieved. The consensus ratings and interpretation of scoring criteria were subsequently reviewed and validated by ET and HDL. Based on observations from this pilot exercise, minor refinements were made to the wording and instructions of several scoring items to improve clarity and consistency of interpretation before commencing the main assessment phase. Subsequently, in the initial testing phase, 98 MR2a reports were independently evaluated by two student assessors (NB and AP). Inter-rater agreement (IRA) per criterion was assessed using weighted Cohen’s kappa in IBM SPSS [47–49]. Agreement patterns were additionally visualized using a Microsoft Excel-generated color-coded matrix (heatmap) [50]. Furthermore, assessment duration was tracked to estimate the practical feasibility of BRANT-MERQS-2A. Further testing and implementation of the BRANT-MERQS-2A In the final testing phase, 61 final-year master’s students in Pharmaceutical Sciences independently applied BRANT-MERQS-2A to the same 98 MR2a reports from the initial testing phase, all authored by other individuals. Participation was voluntary but encouraged. Participation was open to all final year master’s students without additional eligibility criteria. Students were randomly assigned to unique MR2a identification numbers. An allocation key ensured that assessors remained blinded to co-assessors, minimizing potential bias. Scores were submitted via an online Qualtrics survey [51], which also included limited background questions such as prior internship experience (weeks). The full survey is available in the Supplementary Material. To ensure replication, assessments were conducted across six consecutive rounds, so that each MR2a report was evaluated by at least three different students. Inter-rater reliability (IRR) and IRA per criterion were calculated using the Intraclass Correlation Coefficient (ICC) and the Gwet’s AC1 (AC1) [52–55]. ICC values were calculated in IBM SPSS and AC1 values in Microsoft Excel [47, 50]. A two-way random model with absolute agreement was applied for ICC [52]. IRR was interpreted using predefined agreement thresholds based on existing literature [52, 56]. ICC values between 0.50 and 0.75 were considered indicative of moderate agreement according to Koo and Li [52], while Gwet’s AC1 values between 0.41 and 0.60 and between 0.61 and 0.80 were interpreted as moderate and substantial agreement, respectively, based on the benchmarks proposed by Landis and Koch [56]. AC1 evaluates absolute agreement, whereas the ICC assesses the consistency of relative ratings across items [57]. Overall IRR of BRANT-MERQS-2A was subsequently calculated using the ICC (two-way random model, absolute agreement) in Microsoft Excel [50, 52]. Quality of MR2a reports In addition to reliability analysis, the quality of the 98 MR2a reports was evaluated by calculating mean scores for each criterion and the overall scores throughout the different study phases. Data analysis and graphical visualization were performed in Microsoft Excel [50] Ethics approval Ethical approval was obtained from the Local Biomedical Ethics of the University hospital of Antwerp Committee (ref nr ID 6818, 19 January 2026) for the data collection of the students’ MR2a reports. Results BRANT-MERQS-2A From the original list of BRANT-MERQS criteria [45], those applicable to MR2a and the Belgian context were identified. The resulting selection is presented in Table 1. Table 1Overview of the Original BRANT-MERQS criteria, their applicability to MR2a and the Belgian Context of MR2a, the rationale for exclusion and the final inclusion in the BRANT-MERQS-2A tool. ‘ + ’ indicates included into BRANT-MERQS-2A, whereas ‘- ‘ indicates not included into BRANT-MERQS-2AOriginal Quality CriteriaApplicable to MR2aApplicable to Belgian ContextExplanationIntegrated in BRANT-MERQS-2AGeneralExplanation benefit and purpose of MRYesNoNot requested in e-form–General characteristics (birthdate, gender)YesYes + Allergies and intolerancesYesNoPatient records required–Lab valuesNoNoPatient records required–Current pathologiesYesYes + Easy and understandable languageYesNoNot requested in e-form–Quiet and familiar environmentYesNoNot requested in e-form–DrugsOverview chronic medicationYesYes + Overview self-care medicationYesYes + Indications known by patientYesYes + Medication schedule recentYesNoNot requested in e-form–Indications still topicalNoNoPatient records required–EffectivenessNoNoPatient records required–Side effectsYesYes + Drugs treating side effectsYesNoPatient records required–Interactions with other drugs/foodYesYes + Relevance of interactionsYesYes–Correct doseNoNoPatient records required–Vaccination statusYesYes–User-friendly administration formYesNoNot requested in e-form–Control of storage medicationYesNoNot requested in e-form–TreatmentTreatment constructionNoNoPatient records required–Treatment choice as function of comorbiditiesNoNoPatient records required–Drug changes + motivationNoNoPatient records required–Non-medical measuresYesNoNot requested in e-form–UndertreatmentYesYes + OvertreatmentYesYes + Tapering of medicationYesYesOverlap with overtreatment–Addiction riskYesNoPatient records required–AdherenceYesNoNot requested in e-form–Motivation control by healthcare providersNoNoPatient records required–Interpretation lab valuesNoNoPatient records required–Lab values lined to pathologies and drugsNoNoPatient records required–Tools: medication schedule, medication box…YesNoNot requested in e-form–Consultation of GPReport to GPYesNoNot verifiable–Elaboration of specific interventionsYesYesOverlap with Interventions sufficiently reasoned–Interventions sufficiently reasonedYesYes + Intervention plan discussed with patientYesNoNot verifiable–Actions without consulting GP (non-medical)YesYes + Actions after consulting GPYesNoNot verifiable–Follow-up with patientYesNoNot verifiable–Used sources and toolsAvailability of lower cost alternativesYesYesNot requested in e-form–Use of reliable toolsYesNoNot requested in e-form–BibliographyYesNoNot requested in e-form–Reliable literatureYesNoNot requested in e-form– Table 1Overview of the Original BRANT-MERQS criteria, their applicability to MR2a and the Belgian Context of MR2a, the rationale for exclusion and the final inclusion in the BRANT-MERQS-2A tool. ‘ + ’ indicates included into BRANT-MERQS-2A, whereas ‘- ‘ indicates not included into BRANT-MERQS-2AOriginal Quality CriteriaApplicable to MR2aApplicable to Belgian ContextExplanationIntegrated in BRANT-MERQS-2AGeneralExplanation benefit and purpose of MRYesNoNot requested in e-form–General characteristics (birthdate, gender)YesYes + Allergies and intolerancesYesNoPatient records required–Lab valuesNoNoPatient records required–Current pathologiesYesYes + Easy and understandable languageYesNoNot requested in e-form–Quiet and familiar environmentYesNoNot requested in e-form–DrugsOverview chronic medicationYesYes + Overview self-care medicationYesYes + Indications known by patientYesYes + Medication schedule recentYesNoNot requested in e-form–Indications still topicalNoNoPatient records required–EffectivenessNoNoPatient records required–Side effectsYesYes + Drugs treating side effectsYesNoPatient records required–Interactions with other drugs/foodYesYes + Relevance of interactionsYesYes–Correct doseNoNoPatient records required–Vaccination statusYesYes–User-friendly administration formYesNoNot requested in e-form–Control of storage medicationYesNoNot requested in e-form–TreatmentTreatment constructionNoNoPatient records required–Treatment choice as function of comorbiditiesNoNoPatient records required–Drug changes + motivationNoNoPatient records required–Non-medical measuresYesNoNot requested in e-form–UndertreatmentYesYes + OvertreatmentYesYes + Tapering of medicationYesYesOverlap with overtreatment–Addiction riskYesNoPatient records required–AdherenceYesNoNot requested in e-form–Motivation control by healthcare providersNoNoPatient records required–Interpretation lab valuesNoNoPatient records required–Lab values lined to pathologies and drugsNoNoPatient records required–Tools: medication schedule, medication box…YesNoNot requested in e-form–Consultation of GPReport to GPYesNoNot verifiable–Elaboration of specific interventionsYesYesOverlap with Interventions sufficiently reasoned–Interventions sufficiently reasonedYesYes + Intervention plan discussed with patientYesNoNot verifiable–Actions without consulting GP (non-medical)YesYes + Actions after consulting GPYesNoNot verifiable–Follow-up with patientYesNoNot verifiable–Used sources and toolsAvailability of lower cost alternativesYesYesNot requested in e-form–Use of reliable toolsYesNoNot requested in e-form–BibliographyYesNoNot requested in e-form–Reliable literatureYesNoNot requested in e-form– Overview of the Original BRANT-MERQS criteria, their applicability to MR2a and the Belgian Context of MR2a, the rationale for exclusion and the final inclusion in the BRANT-MERQS-2A tool. ‘ + ’ indicates included into BRANT-MERQS-2A, whereas ‘- ‘ indicates not included into BRANT-MERQS-2A The BRANT-MERQS-2A was developed in English, with a Dutch translation provided for the students during the further testing phase. Both versions are available in the Supplementary Material. Inter-rater agreement and inter-rater reliability of the BRANT-MERQS-2A Figure 1 presents the IRA between the two student assessors (NB and AP) for each BRANT-MERQS-2A criterion during the initial testing phase. Detailed, report-level scores are provided in the Supplementary Material.Fig. 1Visualisation of Inter-rater agreement (IRA) as reflected by similarities and differences in BRANT-MERQS-2A scores assigned to MR2a reports in the initial testing phase. For each MR2a report, score discrepancies per BRANT-MERQS-2A quality criterion are displayed. Green cells with a value of “1” indicate identical scores between assessors, whereas disagreements are indicated by a value of “0” and colour-coded according to magnitude of the score difference: yellow for a difference of 1 point, orange for a difference of 2 points and red for a difference of 3 points. The lower panel summarizes the frequency of agreement per criterion. IRA was statistically quantified using Cohen’s weighted kappa Visualisation of Inter-rater agreement (IRA) as reflected by similarities and differences in BRANT-MERQS-2A scores assigned to MR2a reports in the initial testing phase. For each MR2a report, score discrepancies per BRANT-MERQS-2A quality criterion are displayed. Green cells with a value of “1” indicate identical scores between assessors, whereas disagreements are indicated by a value of “0” and colour-coded according to magnitude of the score difference: yellow for a difference of 1 point, orange for a difference of 2 points and red for a difference of 3 points. The lower panel summarizes the frequency of agreement per criterion. IRA was statistically quantified using Cohen’s weighted kappa During the final testing phase, the same 98 MR2a reports were assessed by the student cohort. Figure 2 presents the IRA and IRR for each criterion, including both numerical and visual representations of the ICC and AC1. Agreement ranged from poor to moderate across all criteria, with an overall ICC (two-way random model, absolute agreement) of 0.62, indicating moderate agreement [52]. Detailed, report-level scores are provided in the Supplementary Material.Fig. 2Visualisation of Inter-rater agreement (IRA) based on similarities and differences in BRANT-MERQS-2A scores assigned by three randomly selected student assessors across MR2a reports during the extended testing phase. Green cells (value = 1) indicate complete agreement among all raters. Cells with a value of 0.33 indicate partial agreement, with two raters assessing the same score and one rater differing. Within these cells, yellow denotes a one-point difference, orange a two-point difference and red the maximum possible score difference. Cells with a value of 0 indicate complete disagreement among the three raters. Within these, orange represents a two-point difference between the minimum and maximum score assigned, and red represents the maximum difference of three points. The lower panel summarizes the frequency of agreement per criterion. Percent agreement (%) and Gwet’s AC1 were quantified with Microsoft Excel to present the IRA statistically, whereas the Intraclass Correlation Coefficient (ICC) (two-way random model, absolute agreement) was quantified in SPSS to present the inter-rater reliability statistically Visualisation of Inter-rater agreement (IRA) based on similarities and differences in BRANT-MERQS-2A scores assigned by three randomly selected student assessors across MR2a reports during the extended testing phase. Green cells (value = 1) indicate complete agreement among all raters. Cells with a value of 0.33 indicate partial agreement, with two raters assessing the same score and one rater differing. Within these cells, yellow denotes a one-point difference, orange a two-point difference and red the maximum possible score difference. Cells with a value of 0 indicate complete disagreement among the three raters. Within these, orange represents a two-point difference between the minimum and maximum score assigned, and red represents the maximum difference of three points. The lower panel summarizes the frequency of agreement per criterion. Percent agreement (%) and Gwet’s AC1 were quantified with Microsoft Excel to present the IRA statistically, whereas the Intraclass Correlation Coefficient (ICC) (two-way random model, absolute agreement) was quantified in SPSS to present the inter-rater reliability statistically BRANT-MERQS-2A evaluation time Figure 3 displays the assessment time of the two assessors during the initial testing phase and the mean assessment time spent per MR2a report for the student cohort. Figure 4 illustrates the change in assessment duration between the first and final evaluations per student.Fig. 3Overview of assessment duration for quality control of MR2a reports with BRANT-MERQS-2A. For each MR2a report quality control, time was tracked during the evaluation. For the student cohort, the mean time needed for the assessment across the three students is displayedFig. 4Time difference between the first and final quality assessment of MR2a reports with BRANT-MERQS-2A in the final testing phase. An individual learning effect was observed in 43 out of 61 students (70.5%). For 9 students (14.8%), there was no difference in assessment time observed, while another 9 students (14.8%) showed an increase in assessment time Overview of assessment duration for quality control of MR2a reports with BRANT-MERQS-2A. For each MR2a report quality control, time was tracked during the evaluation. For the student cohort, the mean time needed for the assessment across the three students is displayed Time difference between the first and final quality assessment of MR2a reports with BRANT-MERQS-2A in the final testing phase. An individual learning effect was observed in 43 out of 61 students (70.5%). For 9 students (14.8%), there was no difference in assessment time observed, while another 9 students (14.8%) showed an increase in assessment time The mean assessment time was 18 min and 18 s for rater A and 22 min and 40 s for rater B, yielding an overall mean time of 20 min 29 s. For the student cohort, mean assessment time decreased across rounds, from 43 min and 28 s in the first round to 25 min and 6 s in the final round. Quality of MR2a reports Figure 5 summarizes the mean quality scores per BRANT-MERQS-2A criterion across the 98 evaluated MR2a reports, as assigned by assessor A, assessor B and the student cohort (mean). The figure also includes a boxplot illustrating the distribution of the overall quality scores. The mean overall quality score is 1.61 (54%).Fig. 5Overview of quality scores per BRANT-MERQS-2A criterion for the 98 assessed MR2a reports, together with a boxplot illustrating the distribution of total quality scores. Total report scores were calculated as the mean of the overall scores assigned by the individual assessors for each report Overview of quality scores per BRANT-MERQS-2A criterion for the 98 assessed MR2a reports, together with a boxplot illustrating the distribution of total quality scores. Total report scores were calculated as the mean of the overall scores assigned by the individual assessors for each report Discussion BRANT-MERQS-2A—Framework and interpretation BRANT-MERQS-2A was developed as a practical and reliable quality control framework to support MR2a implementation and promote consistency in professional practice. Scores should be interpreted as relative measures of adherence to predefined quality standards rather than as absolute indicators of the clinical quality or patient benefit. High scores therefore reflect appropriate execution of key procedural elements, while acknowledging that important aspects of clinical decision-making and patient impact extend beyond what is captured by this framework. The tool does not assess appropriateness of clinical decisions, only their documented representation within the MR2a report. The ultimate selection of criteria included in Table 1 was guided by the definition and scope of MR2a, specifically taking into account the type of information available within this level of MR, and excluding elements that require access to more comprehensive patient records as in higher-level reviews. BRANT-MERQS-2A—Reliability and discriminatory capacity Strong IRA during the initial testing phase represents a key strength of BRANT-MERQS-2A. The two student assessors (NB and AP) were intensively involved in the development of the instrument and its accompanying scoring manual, which likely contributed to their shared understanding of MR2a report structure and expectations. Moreover, this in-depth involvement was essential for critically evaluating and refining the manual, thereby ensuring its clarity and usability. Greater variability in the scoring of individual criteria was observed during the further testing phase, with percent agreement (%) ranging from 20 to 39%. At the time of the assignment, students had limited practical experience with MR2a. Although their educational background enabled them to complete the task successfully, this group may lack the additional clinical depth, as present in the initial testing phase, which may account for the lower level of complete agreement observed. Analysis revealed that criteria differ in their discriminatory capacity. Although Overview of Chronic Medications showed moderate apparent agreement (% = 38,78%; AC1 = 0.471) [56], the ICC was very low (ICC = 0.078) [52]. This reflects the ICC’s sensitivity to restricted variance: when most reports receive similar scores, ICC values remain low despite consistent scoring. In this case, the limited variability is expected due to its standardized and largely copy-paste nature of medication list in the e–form. In contrast, Interaction with other Drugs/Food demonstrated a high ICC (ICC = 0.767) [52], indicating that assessors consistently distinguished between higher- and lower-quality reports and assigned comparable absolute scores across reports. This criterion has strong discriminatory capacity, allowing meaningful differentiation between good and poor performances. Identifying and substantiating clinically relevant interactions requires higher-level documented clinical reasoning skills, likely resulting in greater variability between reports and thus effectively capturing true differences in report quality. At the aggregate level, overall agreement remained moderate (ICC = 0.62) [52], indicating that BRANT-MERQS-2A performs more reliably when criteria are considered collectively, as rater-specific variability is averaged across multiple items. Variability between criteria should therefore not be interpreted as a limitation but as a reflection of their inherent differences in discriminatory potential. Previous research indicates that a subset of BRANT-MERQS criteria can approximate overall MR quality, supporting the interpretability of the results despite criterion-level variability [58]. Future research may therefore explore whether reduced sets of criteria are sufficient for reliable assessments with BRANT-MERQS-2A. Professional judgement and future directions Complete uniformity in scoring is neither expected nor desired. As a self- and peer-assessment instrument, BRANT-MERQS-2A aims to stimulate reflection on overall report quality rather than to rigid enforcement. Given the inherently individualized nature of MR2a, complete uniformity in scoring would suggest a purely procedural application, leaving little room for professional judgement and patient-specific adaptation elements that are essential to high quality, patient-centered MR2a services. However, the comprehensive evaluation cannot solely rely on written reports, as these capture procedural consistency rather than full clinical relevance or patient-perceived value. Additional dimensions, such as patient-reported experiences measurements (PREM) and outcome measurements (PROM), should therefore be considered to capture quality and effectiveness of the complete MR2a service [45, 59, 60]. In this context, methodological frameworks such as COSMIN may become particularly relevant in future work focusing on PREMs and PROMs, where psychometric validation is essential [61]. Future research should explore complementary assessment strategies that combine procedural, clinical outcome and experience-oriented measures to evaluate MR quality comprehensively. Currently, scoring relies exclusively on the written MR2a report generated from the e-form. Because the template does not explicitly prompt all BRANT-MERQS-2A elements, relevant clinical reasoning and patient-specific considerations may be insufficiently documented, despite being addressed during the MR2a, contributing to score variability. This highlights the need to better align documentation templates with quality criteria or to complement report-based evaluation with additional information sources. Statistical considerations While the ICC is typically applied to numerical data, it can be used for ordinal scale when category distances are assumed equal [52, 57, 62, 63]. However, ICC values may be artificially low when variance between observations is limited (paradox) [54, 64, 65]. To address this limitation, AC1 was used as a paradox-resistant measure of agreement, especially for standardized criteria [54, 55, 64]. Together, these statistics give complementary insights into rater agreement and rating consistency. Validated agreement benchmarks specific to MR quality assessment tools are currently lacking. Therefore, interpretation of ICC and AC1 values was based on commonly used literature thresholds. Practical feasibility and learning effect A learning effect was observed, with 70.5% of students reducing their assessment time across rounds, indicating increasing experience and familiarity with BRANT-MERQS-2A criteria and their location within the e-form. For some students, assessment time increased, possibly reflecting greater case complexity, although this relationship was not examined in this study. The average reduction in assessment time was modest (approximately 4–7 s per report), and the relatively low R2 (R2 ≈ 0.33 and R2 ≈ 0.39) suggests that assessment duration is influenced by multiple factors, beyond experience alone (e.g. individual, procedural and case-related factors). Nevertheless, the overall downward trend supports the feasibility of BRANT-MERQS-2A in routine practice. Notably, self-assessment may require less time than peer-assessment (as conducted here) due to prior familiarity with the content and context. Quality of MR2a reports As presented in Fig. 5, Overview of chronic medications received the highest scores, likely due to automatically generated medication lists in the e-form, whereas Interactions with other drugs/food scored lowest, reflecting their greater clinical complexity. The mean overall report quality score of the 98 MR2a reports is 1.61 (54%), indicating moderate report quality, which suggests room for improvement while showing that most reports were not of very low quality. Use of BRANT-MERQS-2A as a self-assessment tool may help pharmacists and students structure the MR2a process and improve report quality. Strengths and limitations BRANT-MERQS-2A represents, to our knowledge, the first attempt to objectively assess the quality of MR2a reports. This tool provides a practical approach for pharmacists to evaluate their own reports or those of colleagues and may support communication of service quality to other healthcare professionals and stakeholders. However, BRANT-MERQS-2A currently evaluates only written reports and does not capture other key components, such as patient consultation and follow-up. This highlights the need for further research to develop complementary tools and to evaluate the applicability and reliability of BRANT-MERQS-2A in different settings and user populations. Even in its current form, BRANT-MERQS-2A offers a structured framework that is preferable to the absence of any objective assessment method. An additional limitation is that both the MR2a reports and assessments were performed by pharmacy students rather than practicing pharmacists, which may limit applicability to routine professional practice. However, as MR2a and standardized MR2a quality assessment remain relatively novel concepts, familiarity with these processes is currently also limited among practicing pharmacists. Moreover, final-year students had received specific MR2a training and conducted the reviews during supervised internships. Clinical judgement is not directly assessed at this moment, only its documentation within the MR2a report is evaluated. This focus on the written report reflects an initial step in a broader evaluation framework, with further development needed to fully capture clinical decision-making, patient-centered aspects of MR2a and ultimately to enable assessment of clinical outcomes. Conclusion BRANT-MERQS-2A provides a feasible and structured approach to assessing quality of MR2a reports, demonstrating moderate inter-rater reliability and practical usability. While variability at the criterion level reflects the inherent complexity of MR2a, the tool shows potential for self- and peer assessment and offers a valuable foundation for further refinement and integration into routine quality monitoring. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 2428 KB)Supplementary file2 (PDF 71 KB) Supplementary file1 (DOCX 2428 KB) Supplementary file2 (PDF 71 KB)