Recent demographic trends project a continued increase in the number of older adults [1]. This is closely linked to a growing global prevalence of chronic diseases [2], often accompanied by complex challenges, such as multimorbidity [3] and polypharmacy [4]. Polypharmacy, commonly defined as the concurrent use of 5 or more medications, is associated with various adverse outcomes [5,6], such as unwanted drug-drug interactions (DDIs), potentially leading to adverse drug events [7]. DDIs are even more likely if multiple providers are involved in prescribing medications to patients with polypharmacy [8]. For instance, a German study found that only 34.7% of patients treated with prescription medications receive their medications from a single physician [7], underscoring the challenges of managing chronic diseases.

A key challenge in managing polypharmacy is the health care providers’ lack of comprehensive knowledge of the patients’ current medications. This knowledge gap compromises both the safety and effectiveness of pharmacological treatment. To address this, the implementation of comprehensive medication plans (MPs) encompassing prescribed and over-the-counter (OTC) medications in a structured and accessible format for both patients and providers has been widely recommended [9,10]. Empirical evidence supports the positive impact of MPs on medication safety. For example, the use of MPs has been associated with reduced hospitalization rates and fewer incidents of medication misuse [11], as well as a slight decrease in negative DDIs [12].

Despite these benefits, the adoption and consistent use of MPs is still insufficient [7,13-15]. A major barrier is the frequent incompleteness of MPs. For example, a study conducted in Germany found that, in a regional sample, only 6.3% of the nationwide MPs met all formal requirements, including active substance, dosage, or reasoning [9]. Moreover, evaluations of the MP in Germany reveal discrepancies between documented and actual medication use in approximately 78% of cases, particularly regarding OTC medication and dietary supplements [16].

Another intervention promising improvement in medication therapy management is electronic prescriptions (e-prescriptions). They offer a way to significantly contribute to improvements, such as reducing the risk of medication errors and adverse events as well as improving efficiency in medication management if well designed [17,18]. The full potential of MPs and e-prescriptions can be realized when integrated into clinical decision support systems (CDSS). The benefits of CDSS in clinical practice have already been demonstrated in a number of disease conditions [19-22], and their application in medication safety is particularly promising. For example, CDSS can identify DDIs overlooked by physicians, thereby improving the quality of care [22]. Furthermore, CDSS may prompt physicians to engage in more critical and reflective prescribing practices [23,24].

The objective of the eRIKA study (e-prescription as an element of interprofessional care pathways for continuous medication therapy management [eRezept als Element interprofessioneller Versorgungspfade für kontinuierliche AMTS]) is to improve the safety of medication therapy through the combined use of e-prescriptions, claims data, MPs, CDSS, and interprofessional communication. To comprehensively evaluate the intervention, the study aims to investigate (1) its effectiveness as well as (2) health-economic and (3) implementation-related aspects.

This study protocol was written in accordance with the TIDieR (Template for Intervention Description and Replication) reporting checklist (Checklist 1) [25]. To comprehensively evaluate this complex intervention, a mixed perspective, multiphase study design will be employed. Phase 1 is designed as a cluster randomized controlled trial (c-RCT) with a relatively small sample size to allow for initial evaluation under controlled conditions. This is followed by a second phase involving a larger-scale quasi-experimental study with a matched control group (CG), scaling up the intervention and enabling the examination of the intervention’s effectiveness, cost-effectiveness, and implementation-related aspects in real-world settings.

The eRIKA intervention is designed to guarantee that for every patient receiving pharmacotherapy, an up-to-date and comprehensive electronic MP is generated, updated, and centrally stored, which can be accessed by patients, treating outpatient physicians, and pharmacists. Furthermore, medical information on the patient is electronically extracted from claims data provided by statutory health insurance (SHI) funds and is used to inform outpatient physicians about diagnoses, diagnostic procedures, and treatment history to date.

The eRIKA intervention is tested and implemented in participating outpatient practices and pharmacies. It is a digital software solution that integrates (1) the use of claims data in clinical practice, (2) the generation of an MP, (3) CDSS, (4) e-prescriptions, and (5) interprofessional communication.

Claims data from the SHI funds concerning prior prescriptions and selected clinically relevant information (such as diagnoses, medication prescription history, performed diagnostic and therapeutic procedures, treating outpatient physicians, hospital admissions and treatment, remedies, and aids) are fed into the software to ensure that MPs are complete and up to date at all times. This provides outpatient physicians with a comprehensive overview of the patient’s medical and medication history, which enables more efficient and safer medication prescriptions. The provision of medical and pharmaceutical information ensures that pharmacists have an overview of the patient’s current and past medications during their pharmaceutical consultations.

By using e-prescriptions via the eRIKA software and claims data, an MP is prefilled to be checked and manually supplemented with OTC medication, thus providing an up-to-date overview of the current medication regimens for use in outpatient practices and pharmacies. Furthermore, this MP is made accessible to patients, providing them with comprehensive information concerning their medication regimen.

Regarding CDSS, the eRIKA software also provides information and alerts for avoidable risks in medication therapy, for example, on contraindications, dosing errors, potentially inadequate medication for age, medication interactions, Dear-Doctor letters (Rote-Hand-Briefe), as well as guidance based on the guidelines for drug therapy in multimorbidity (Leitlinie Arzneimitteltherapie bei Multimorbidität), all of which are relevant to prescription decisions. Alerts are displayed in real time during prescription entry by the physician and are triggered automatically when rule-based checks detect a mismatch between the prescribed medication and patient-specific characteristics such as age, diagnoses, or concomitant medication. Alerts are categorized by severity (levels 1‐3) and can be addressed by modifying or canceling the prescription. Alternatively, physicians may override alerts if they consider the prescription clinically appropriate. For high-severity alerts (levels 2 and 3), physicians are required to actively acknowledge the alert and provide a brief justification, whereas lower-severity alerts may be acknowledged without justification. All alerts, physician responses, and overrides are automatically documented within the system, enabling assessment of alert exposure and handling during the intervention. A tailored subset of alerts relevant for pharmaceutical counseling is presented to pharmacists.

e-Prescriptions via the eRIKA software are issued in accordance with the legal requirements under German law.

Regarding interprofessional communication, the software further enables the outpatient physician to document remarks regarding the selected medication and assign corresponding levels of urgency.

An overview of the structured eRIKA process is provided in Figure 1. Following patient consent (1), outpatient physicians use the eRIKA software during routine consultations to access claims-based medical information and the centrally stored MP. After conducting the anamnesis and reviewing the current medication (2.1), supported by CDSS alerts (2.2), new medications are prescribed electronically (2.3). Issued e-prescriptions and imported claims data automatically update the server-side MP (3.1). Furthermore, an up-to-date MP is provided to the patient as a printout (3.2). Next, patients redeem their e-prescriptions in participating pharmacies (4). In this step, pharmacists access the MP and relevant patient information (5.1) and add OTC medications and dietary supplements to the MP if necessary (5.2). The pharmacist-led structured pharmacotherapy safety check is supported by a tailored set of alerts relevant to pharmaceutical counseling (5.3). Pharmacists provide medication counseling (5.4) and, if necessary, contact the treating physician (electronically or via telephone) regarding medication-related remarks or substitution proposals for review (5.5). Following optional physician review, the MP is updated accordingly and made available to all involved health care professionals via the eRIKA software as the server is updated automatically (6), as well as to patients in printed form (7).

Reimbursement is tied to the completion of defined process steps to support adherence to the intended process. Physicians had to fulfill the following five steps: patient enrollment (1); conducting the anamnesis (2.1) and a structured safety check (2.2) and issuing e-prescriptions (2.3); providing an updated MP (3.1 and 3.2); pharmacists dispensing at least 1 e-prescription (4); and reviewing the corresponding risk analysis (5.1), documenting any required modifications to the MP (5.3), and issuing an updated MP if necessary (7).

Physicians and pharmacists received initial training sessions during which the software developer explained the system’s functionality, providing opportunities to ask questions and report technical issues. In addition, a brief process manual was integrated directly into the software for ongoing reference.

For the evaluation of the primary outcome, 10 clusters of outpatient practices and pharmacies in geographical proximity are randomized into 2 groups (5 IG and 5 CG). Because, to our knowledge, no comparable studies are available to support plausible assumptions regarding the primary outcome, a moderate Cohen d effect size of 0.5 is assumed, in line with established conventions for studies with novel outcomes [28,29]. The assumed intracluster correlation coefficient (ICC) of 0.05 is based on the evidence from c-RCTs conducted in similar primary care and outpatient settings, where ICCs typically range from 0.01 to 0.10 [30-32]. Given these assumptions, an α-error of .05 (2-sided), and a power of 0.8, the overall sample size is 138 (2×69). Based on experience from prior studies in comparable outpatient and primary care settings [33,34], as well as to account for possible patient withdrawal and loss of analyzable primary outcome data due to incomplete or manually transmitted MPs, a dropout rate of 40% is expected, increasing the sample size to 230 patients (2×115). The resulting number of patients per practice is 23.

For the outcome evaluation, the incidence of all-cause hospitalization or all-cause mortality (primary outcome) for patients with polypharmacy is estimated to be 32.9%, based on a preliminary analysis of BARMER claims data. Given the inconsistent evidence regarding the impact of medication management interventions on outcomes such as hospitalization and mortality [11,33,41-45] and the fact that these outcomes are influenced by multiple clinical and contextual factors beyond the intervention itself, a conservative absolute risk reduction of 4 percentage points was assumed, corresponding to a reduction to 28.9%. The evaluation is carried out for 2 groups, each in comparison to a matched CG: for patients with (1) 3 to 4 medications and (2) 5 or more medications. With an α-error of .05 and a power of 1-β=0.80, based on a 2-sided Fisher exact test (for binary outcomes), this yields a required sample size of 1411 patients per group, resulting in a total of 2822 patients. Accounting for a dropout rate of 20%, the required sample size is adjusted to 3528 participants.

Drawing on experience from previous projects, an average of 10 to 20 patients per practice is expected, corresponding to approximately 175 to 350 outpatient physicians and 116 to 233 outpatient practices and corresponding pharmacies.

The IG is assigned a data-derived CG by selecting and matching comparable insured individuals from the claims data provided by the participating SHI funds. These individuals meet the study’s inclusion criteria and have not received treatment in any of the participating outpatient practices or pharmacies during the intervention phase. An iterative matching approach is employed, consisting of a primary procedure and alternative procedures. If the primary method of direct matching based on relevant matching criteria (eg, age, gender, and prescribed medications) does not yield a sufficiently large and balanced matched CG, an alternative procedure based on propensity score matching is applied, using relevant matching criteria as covariates. This approach aligns with standard practice in the application of matching techniques [46], as there is no single optimal choice of matching approach for any given analysis [47]. To assess residual confounding after matching, covariate balance between the IG and the matched CG is evaluated using standardized mean differences and visual diagnostics for all matching variables and key preintervention characteristics (eg, health care use and comorbidity indicators) [48,49]. To address any residual imbalance, outcome analyses are conducted using regression models adjusted for baseline covariates that remain imbalanced after matching. As a sensitivity analysis, alternative weighting approaches are applied to assess the robustness of the results to the choice of matching method [50].

For phase 1, ethical approval was granted by the ethics committee of the Medical Association of Westphalia-Lippe (2023‐526-f-S). For phase 2, ethical approval was granted by the ethics committee of the Medical Association of Westphalia-Lippe (2024‐599-f-S) and the Medical Association of Saarland (208/24).

This study is conducted in accordance with the Declaration of Helsinki. The CG in phase 2 is derived from secondary pseudonymized claims data. Informed consent from CG individuals is not obtained. However, the study poses minimal risk, and a waiver of informed consent for the use of pseudonymized secondary data was reviewed and approved by the responsible ethics committee as well as regulatory authorities. All data are collected, pseudonymized, and stored in accordance with German General Data Protection Regulation legislation based on the European General Data Protection Regulation.

Patients in phase 1 and IG patients in phase 2 voluntarily provide written informed consent prior to participation. Participants are informed about the project, the evaluation procedures (including data linkage of MP, software data, and claims data), participation requirements, and their rights. In phase 2, patients additionally consent to participate in interviews. By signing a separate consent form, they agree to be contacted by research staff. No financial compensation is provided to patients in the study.

Participating outpatient practices and pharmacies consent to participate through their project participation agreement, replacing the need for individual staff consent forms. Study staff are informed about study procedures prior to each data collection activity. By completing questionnaires or voluntarily participating in interviews or focus groups, staff consent to the project-related processing, use, and storage of their data.

All participants may withdraw their informed consent at any time without providing reasons and without facing any disadvantages. In the event of withdrawal, no further data will be collected, and previously collected data will be anonymized in accordance with data protection regulations.

Study data are stored in encrypted form with access restricted to authorized project staff. Personal contact data are stored separately from research data and deleted in accordance with data protection regulations.

The protocol was registered with the German Clinical Trials Register database (phase 1: DRKS00032777 and phase 2: DRKS00035030).

There is an urgent need to improve the safety of medication therapy for patients with polypharmacy. The growing number of older adults is accompanied by rising multimorbidity and polypharmacy [1-4], increasing the risk of DDIs [5-7], particularly when multiple providers are involved [8]. Meanwhile, MPs often remain incomplete or underutilized [7,13-15]. Structured MPs and CDSS have shown positive effects on medication therapy safety, such as reducing DDIs, hospitalizations, and inappropriate medication use [11,12]. However, the widespread implementation of MPs remains limited, highlighting the urgent need for better integration into routine care. The eRIKA software is a promising intervention to improve medication therapy safety for patients with polypharmacy. The combined use of e-prescriptions, claims data, automatically generated MPs, and CDSS has the potential to provide health care providers with an up-to-date overview of the current medication, contribute to the completeness and actuality of the available MP, and support outpatient physicians and pharmacists in their treatment or consultation process. The eRIKA intervention extends forms of care that have already been tested and evaluated in other projects such as the AdAM (Anwendung für digital unterstütztes Arzneimitteltherapie-Management) [33] or TOP (Transsektorale Optimierung der Patientensicherheit) [34] project by adding the use of e-prescriptions and emphasizing the essential role of pharmacists.

A key strength of the study is the alignment of the eRIKA project with the Medical Research Council (MRC) framework for the development and evaluation of complex interventions [55]. This includes a comprehensive approach encompassing intervention development, implementation, and multilevel evaluation, including formative components. Furthermore, the study is structured in 2 phases to ensure both rigor and scalability. Phase 1 is conducted as a c-RCT, allowing for a controlled assessment of the intervention’s feasibility within a clinical setting. In phase 2, the study involves a larger sample size to evaluate the effectiveness and large-scale implementation. Moreover, with this comprehensive analysis approach, a multitude of relevant outcomes for the effectiveness of the eRIKA software and process are addressed, as recommended by the literature [56]. The intervention itself is intersectoral in nature, involving coordinated efforts across different sectors of the health care system, which enhances its practical relevance.

The study faces limitations that may affect the interpretation and implementation of its findings. In phase 1, the small number of clusters may reduce the generalizability of the results. In addition, the primary outcome in phase 1 (congruence between MP and claims data) is novel, and therefore the underlying effect size and ICC assumptions are based on methodological conventions rather than outcome-specific empirical estimates. Furthermore, claims data are subject to delays and omissions and may not fully reflect actual medication use. Observed discrepancies between the MP and claims data may therefore result from timing differences rather than true inconsistencies, which should be considered when interpreting the congruence outcome in phase 1. Phase 2 is designed as a quasi-experimental study, which, while useful for evaluating interventions in more naturalistic settings, is not exempt from potential confounding and bias. Rigorous (matching) methods are necessary to mitigate such risks. Additionally, the implementation of the intersectoral interventions may present practical challenges, including reliance on service providers, variations in stakeholder engagement, coordination between sectors, technical challenges (at the point of installation or software usage), and contextual factors that could influence intervention fidelity and outcomes. Particularly the implementation of other digital interventions on a national level, such as e-prescriptions in general or the national rollout of electronic patient records, might hinder physicians’ recruitment efforts and intervention fidelity. These factors will be considered in the formative evaluation as well as in the statistical analysis by using a time variable.

Improving medication safety for patients with polypharmacy remains a critical health care priority in Germany. This study provides a comprehensive, multiperspective evaluation of the eRIKA intervention, delivering robust evidence to support its nationwide implementation.