Accessibility settings

This paper is in the following e-collection/theme issue:

Digital Health Reviews (1864) Clinical Information and Decision Making (2128) Digital Health, Telehealth and e-Innovation in Clinical Settings (597) Natural Language Processing (984) Decision Support for Health Professionals (1676) Safety and Error Prevention in Health (268) Design and Usability of Clinical Software and EHRs (379) Machine Learning (2289)

Published on in Vol 28 (2026)

This is a member publication of University of Toronto

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/86760, first published .
AI-Based Automation for Medication Reconciliation: Scoping Review

AI-Based Automation for Medication Reconciliation: Scoping Review

AI-Based Automation for Medication Reconciliation: Scoping Review

Authors of this article:

Juan Pablo Tabja Bortesi1, 2 Author Orcid Image ;   Maria P Becerra1 Author Orcid Image ;   Jonathan Ranisau1 Author Orcid Image ;   Bonnie Wen3 Author Orcid Image ;   Praveen Nadesan4 Author Orcid Image ;   P J Devereaux4, 5, 6 Author Orcid Image ;   Michael McGillion5, 7 Author Orcid Image ;   Jeremy Petch1, 5, 6, 8 Author Orcid Image
Article Authors Cited by Tweetations Metrics

    1Centre for Data Science and Digital Health, Hamilton Health Sciences, Hamilton, ON, Canada

    2School of Epidemiology and Public Health, University of Ottawa, Ottawa, ON, Canada

    3Faculty of Health Sciences, McMaster University, Hamilton, ON, Canada

    4Michael G. DeGroote School of Medicine, Faculty of Health Sciences, McMaster University, Hamilton, ON, Canada

    5Population Health Research Institute, Hamilton, ON, Canada

    6Division of Cardiology, Department of Medicine, McMaster University, Hamilton, ON, Canada

    7School of Nursing, Faculty of Health Sciences, McMaster University, Hamilton, ON, Canada

    8Institute for Health Policy, Management and Evaluation, University of Toronto, 155 College St 4th Floor, Toronto, ON, Canada

    Corresponding Author:

    Jeremy Petch, PhD


    Background: Medication reconciliation (MedRec) has the potential to improve patient safety by enhancing the continuity of medication information across settings. MedRec involves 3 core tasks: the creation of a best possible medication history, the identification of medication discrepancies among medication lists, and the resolution of medication discrepancies. While artificial intelligence (AI) has the potential to improve MedRec, existing reviews have not identified the ways in which researchers have used AI to facilitate MedRec tasks and their constituent subtasks or the level of automation achieved.

    Objective: This scoping review aimed to map how previous research has applied AI to MedRec tasks and subtasks and assess the extent of automation achieved.

    Methods: We searched MEDLINE, Embase, Web of Science, IEEE Xplore, and Compendex in June 2024 for studies that used AI to support a MedRec task or subtask, excluding entirely rule-based tools or studies focused on other aspects of medication management. After screening 2345 unique records, we conducted backward citation searching of studies included at the full-text stage, identifying an additional 795 unique records. We used a 4-stage model of human information processing as a structural lens to guide our considerations of automation, mapping the core tasks of MedRec onto this model.

    Results: A total of 94 studies met the inclusion criteria. All studies addressed subtasks related to the creation of a best possible medication history. Only 2.1% (n=2) of the studies also addressed the identification of discrepancies. Thus, the highest stage of automated information processing achieved was information analysis, although most studies (92/94, 97.9%) only automated information acquisition steps. Most studies (67/94, 71.3%) used free-text clinical notes from the electronic health record, although a significant proportion (21/94, 22.3%) used images of pills or images of other medication-related items. Studies using text-based data used a variety of machine learning methods (eg, recurrent neural networks, conditional random fields, support vector machines, and transformers), whereas those that leveraged images typically used convolutional neural networks. Most studies (61/94, 64.9%) used publicly available data from benchmarking datasets (eg, n2c2 2022) and were strictly model development studies, with only 1.1% (n=1) being usability studies.

    Conclusions: This is the first review to consider the role of AI in the automation of MedRec tasks, offering a basis for prioritizing future development efforts. Current applications of AI to automate MedRec tasks are preliminary, with most work focusing on the extraction of medication information and limited to proof-of-concept model development. Future work should consider addressing infrastructural barriers to the AI-based automation of MedRec tasks (eg, data incompleteness in sources of medication information) and exploring approaches to automate discrepancy resolution. Beyond developing models, there is also a need to implement them in tools and evaluate them in real-world contexts.

    Trial Registration: OSF Registries osf.io/n64u3; https://osf.io/n64u3

    J Med Internet Res 2026;28:e86760

    doi:10.2196/86760

    Keywords

    medication reconciliation; artificial intelligence; AI; health IT; machine learning; patient safety; continuity of care 


    There is a potential for error at every stage of the medication use cycle, which can often lead to serious harms for patients. This avoidable harm is substantial: across care settings, a substantial proportion of adverse drug events (ADEs) have been estimated to be preventable, although estimates vary by context and definition [1-3]. The economic impact of ADEs is considerable, with estimates exceeding hundreds of millions of dollars per year across acute and primary health care systems [4]. Given the scale of the problem, in 2017, the World Health Organization announced Medication Without Harm as its third Global Patient Safety Challenge, outlining strategies to reduce avoidable medication-related harm [5]. One of the chief causes of medication errors is inaccuracies in medication lists used throughout a patient’s care.

    Medication reconciliation (MedRec) is an important procedure for identifying such inaccuracies. MedRec generally includes the creation of a best possible medication history (BPMH) that is compared to a list of medications ordered by a health care professional (eg, a physician, nurse, or pharmacist) to address unintended discrepancies [6]. This typically occurs during care transitions, such as hospital admission and discharge, when medication errors frequently occur [7-10]. MedRec can also be performed in primary care or outpatient settings, often with patients at risk of readmission following discharge [11]. Recognizing the complexity of care transitions, organizations such as Accreditation Canada and the National Institute for Health and Care Excellence recommend enhanced MedRec to improve medication safety [12-15].

    Previous research suggests that MedRec can be effective in identifying medication discrepancies, although evidence of improvements in clinical or health care use outcomes is mixed. A systematic review on the effectiveness of hospital-based, pharmacist-led MedRec found that MedRec was associated with significant reductions in all-cause hospital readmissions and all-cause emergency department visits despite heterogeneity across studies [16]. In contrast, a similar review found no significant reduction in health care use [17]. A Cochrane review evaluating the effect of MedRec on medication discrepancies, patient‐related outcomes, and health care use concluded that the certainty of the evidence in favor of MedRec was very low, noting heterogeneity across studies in intervention implementation, outcome definitions, and intervention effects [18]. The complexity and need for multiple coordinated components in MedRec interventions may explain their inconsistent outcomes [19,20]. These workflows are often time- and labor-intensive [10,21], which can compromise fidelity.

    There have been efforts to develop electronic MedRec tools over the last 2 decades. These tools have various functionalities that may reduce the time and labor demands of traditional, paper-based MedRec, such as interfaces that visually facilitate comparison of medication lists [22] and automated retrieval of community medication lists [23]. Other efforts have focused on methods for conducting MedRec tasks remotely, such as creating an interactive web form or leveraging video visits [24-26]. Due to the need for repetitive data entry, MedRec workflows are ideal targets for artificial intelligence (AI)–driven automation [27]. Although automation has not been the focus of previous literature reviews of MedRec tools, Marien et al [28] conducted a systematic review of electronic MedRec tools and found that only half of the identified tools provided automated discrepancy detection. Ciudad-Gutiérrez et al [29] reviewed a similar set of electronic MedRec tools, noting that approximately half could automatically retrieve medication information. While both functionalities involve important processes that could be automated, no review has yet considered the role of AI in automating MedRec. AI has recently been applied to specific subtasks within the broader workflows of MedRec (eg, extracting medications from free-text notes [30]), but previous literature reviews of MedRec tools have not explored how AI has been leveraged to facilitate MedRec tasks and subtasks. Existing reviews have also examined the use of AI for other aspects of medication use and safety, such as ADE prediction and detection [31], or more broadly for medication management [32,33], but none have focused specifically on MedRec and its constituent tasks and subtasks. As medication errors remain a leading cause of preventable patient harm, mapping the MedRec processes to which AI has been applied is an important step toward understanding its role in facilitating MedRec automation. This is the first review to examine the use of AI for the automation of MedRec processes broken down by specific tasks and subtasks.

    The objective of this scoping review was to map the existing literature on the use of AI for MedRec tasks and subtasks, with a focus on assessing the types of functions automated by the models or tools.


    We conducted this scoping review following guidance from the Joanna Briggs Institute [34]. This review is reported in accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews; Checklist 1) [35].

    Eligibility Criteria

    Following our protocol [36], eligible studies described applications of AI for a task related to MedRec. A previous review of electronic tools that support MedRec described MedRec as consisting of three core tasks: (1) creation of a BPMH, (2) identification of discrepancies among medication lists, and (3) resolution of medication discrepancies [28]. On the basis of these criteria, we included studies that described the development or evaluation of an AI model or AI-based tool that facilitated any of these tasks or a subtask associated with any of these tasks. To maintain relevance, we ensured that the primary objectives of the included studies aligned with MedRec tasks or subtasks.

    Previous scoping reviews have mapped the use of AI or machine learning (ML) for purposes related to medication adherence and ADEs [31,37]. Thus, to avoid duplicating existing work, we excluded studies focusing on adherence, side effects, adverse drug reactions, or allergies, narrowing our review to the core tasks of MedRec. Similarly, we excluded studies restricted to specific medication classes to ensure broader applicability. We also excluded models or tools that were solely rule-based, instead focusing on those that were ML-based or incorporated both ML and rule-based methods.

    We included studies involving participants of any population, as well as those focusing on nonhuman elements such as medication lists. We excluded studies dealing specifically with animals or veterinary medicine. We considered all health care settings (eg, inpatient, outpatient, and primary care). We included journal articles and certain types of gray literature, namely, conference papers and preprints, but excluded abstract-only reports. We considered structured reviews for inclusion but excluded other secondary literature.

    Information Sources, Search Strategy, and Study Selection

    We undertook an initial preliminary search of MEDLINE (PubMed) to identify relevant articles. We used text words used in the titles and abstracts of these articles, as well as index terms used to describe them, to develop the full search strategy (Multimedia Appendix 1), which we adapted for each database. For the AI or ML component of our search, we adapted the AI- or ML-relevant terms from the search strategy used by Syrowatka et al [31] in their review of AI for ADE applications. The databases we searched were MEDLINE (Ovid), Embase (Ovid), Web of Science Core Collection, IEEE Xplore, and Compendex (Engineering Village). We searched the databases from inception to June 10, 2024, without any restrictions. In accordance with our eligibility criteria, we used a filter to exclude records indexed as involving animals and not humans. We also performed backward citation searching based on the articles included at the full-text stage using Citationchaser [38]. Due to resource constraints, we only considered English-language reports. Additional search details are presented in the PRISMA-S (Preferred Reporting Items for Systematic Reviews and Meta-Analyses literature search extension) [39] checklist (Checklist 2).

    After completing the search, we uploaded all records retrieved into Rayyan (Qatar Computing Research Institute). One reviewer manually inspected all records flagged by Rayyan as potential duplicates for deduplication. Following deduplication, at least 2 independent reviewers performed title and abstract screening for each record. Where screeners disagreed about study inclusion, they discussed the records to reach consensus, consulting an additional reviewer if necessary. We retrieved the full texts of potentially relevant records and assessed them in the same manner.

    Data Extraction

    We abstracted data using an extraction tool that we had piloted with 20% of the included studies (Multimedia Appendix 2 [30,40-132]). At least 2 independent reviewers extracted data for each study. Where discrepancies in data extraction arose, the extractors resolved them through discussion to reach consensus, consulting a third reviewer when consensus could not be achieved. We extracted information such as primary objectives, data sources, participant characteristics, training and evaluation set sizes, MedRec tasks and subtasks facilitated by the model or tool, ML methods used, and performance metrics for the best-performing model.

    Data Analysis

    Previous research has explored the concept of automation from various perspectives. Our analytic approach was guided by the 4-stage model of human information processing by Parasuraman et al [133], which outlines system functions that can be automated, including information acquisition, information analysis, decision selection, and action implementation. We mapped the 3 core MedRec tasks of gathering medication data for the creation of a BPMH, identifying discrepancies, and resolving discrepancies (ie, evaluating discrepancies to determine actions) onto the first 3 stages of this model: information acquisition, information analysis, and decision selection, respectively (Figure 1 [28]). The fourth stage, action implementation, extends beyond the scope of core MedRec tasks as defined by Marien et al [28] but may include updating medication lists, communicating changes to the care team, or educating patients. This mapping approach provided a structured lens for assessing how AI-enabled automation could influence MedRec workflows.

    ‎
    Figure 1. Stages of human information processing and corresponding medication reconciliation (MedRec) tasks (in an inpatient admission context) outlined by Marien et al [28]. BPMH: best possible medication history.

    Review Findings

    We identified 2345 records through database searching and an additional 795 unique records from citation searching, ultimately including 94 studies (Figure 2).

    ‎
    Figure 2. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram showing the study selection process. ADE: adverse drug event; MedRec: medication reconciliation; ML: machine learning.

    We found that the included studies were largely restricted to 1 MedRec subtask, primarily using clinical notes in the electronic health record (EHR), although they used a variety of ML methods.

    MedRec Tasks and Subtasks

    All the included studies (94/94, 100%) focused on subtasks related to the creation of a BPMH, whereas 2.1% (2/94) also addressed the identification of discrepancies (Table 1). In terms of BPMH creation subtasks, most studies (92/94, 97.9%) extracted medication information from textual or visual sources [30,40-130]. Only 2.1% (2/94) of the studies focused on verifying the accuracy of already compiled medication information, and both of these studies leveraged a collaborative filtering approach to identify potential omissions from pharmacy medication lists [131,132]. Of the 2 studies that addressed the identification of discrepancies, both identified medications that were consistent as well as inconsistent between medication lists. One study used a conditional random field (CRF) model to extract medication entities from free-text clinical notes and compared them to medications extracted from structured discharge prescriptions using string matching, term co-occurrence, and drug synonyms, labeling medications from the clinical notes as “matched” or “discrepant” [80]. The other study developed a web-based application that retrieved medication lists from the EHR, allowing patients to manually confirm the extracted medications and add missing medications. Differences between the AI-extracted medications and patient-entered medications were flagged by the application [84]. No studies addressed the resolution of discrepancies between medication lists. As such, the highest stage of information processing automation achieved by the included models and tools was information analysis, with most studies (92/94, 97.9%) only automating information acquisition; none automated decision selection (Figure 3).

    Table 1. Characteristics of the included studies (N=94).
    CharacteristicStudies, n (%)
    Year of publication
    2001‐201014 (14.9)
    2011‐202045 (47.9)
    2021‐202435 (37.2)
    Region
    Africa1 (1.1)
    Asia19 (20.2)
    Australia4 (4.4)
    Europe15 (16)
    North America55 (60.4)
    Medication reconciliation taska
    Creation of a BPMHb94 (100)
    Identification of discrepancies2 (2.1)
    Resolution of discrepancies0 (0)
    MLc methoda
    Conditional random field36 (38.3)
    Recurrent neural network34 (36.2)
    Convolutional neural network31 (33)
    Transformer25 (26.6)
    Support vector machine20 (21.3)
    Maximum entropy8 (8.5)
    Random forest6 (6.4)
    K-nearest neighbor6 (6.4)
    Naive Bayes4 (4.3)
    K-means clustering4 (4.3)
    Other21 (22.3)
    Data sourcea
    Clinical note in EHRd67 (71.3)
    User-generated image21 (22.3)
    Pharmacy records3 (3.2)
    Patient-physician conversation transcript2 (2.1)
    Structured prescription list in the EHR1 (1.1)
    User-generated note1 (1.1)
    User-generated structured data1 (1.1)
    Medical internet forum1 (1.1)
    Data collectiona
    Author-collected data33 (35.1)
    i2b2 2009 dataset19 (20.2)
    n2c2 2018 dataset16 (17)
    n2c2 2022 dataset12 (12.8)
    MADE 1.0 2018 dataset7 (7.4)
    NIHe NLMf Pill Image Recognition dataset4 (4.3)
    Other19 (20.2)
    Stage of implementation
    Model development, retrospective93 (98.9)
    Usability study1 (1.1)

    aTotals greater than 100%.

    bBPMH: best possible medication history.

    cML: machine learning.

    dEHR: electronic health record.

    eNIH: National Institutes of Health.

    fNLM: National Library of Medicine.

    ‎
    Figure 3. Conceptual mapping of medication reconciliation (MedRec) tasks and subtasks identified in this review onto the stages of human information processing and frequency of included studies by year. Subtasks are listed as bullet points. BPMH: best possible medication history; NIH: National Institutes of Health; NLM: National Library of Medicine.
    ML Methods

    The included studies used a variety of ML methods, with many using multiple methods (Figure 4). The most commonly used methods included CRFs (36/94, 38.3%) [30,41,43-45,48,52,55,57,67,68,70,71,73,75,78-81,94,95,99,100,103,109,115-117,119,123,126,127], recurrent neural networks (34/94, 36.2%) [30,40,41,45,48-52,59,61,67,68,70-74,76,78,81,85,91,92,94,95,101,103,109,119,123,125-127], convolutional neural networks (31/94, 33%) [40,42,45,49,52,59,64,72,74,76,77,81,82,87,90,92,95-97,101,102,108,112,114,119,120,122-124,127,130], transformers (25/94, 26.6%) [30,46,47,57,60,74,75,85-87,92-95,101,103-107,113,118,119,121,128], and support vector machines (20/94, 21.3%) [53,55,56,59,65,66,73,79,83,86,91,99,100,111,115,123,125-127,130]. Studies using free-text data (eg, notes in the EHR) typically used CRFs, recurrent neural networks, support vector machines, and transformers, whereas image-based studies mainly used convolutional neural networks. Owing to heterogeneity in task definitions (eg, differences in information extraction objectives and input modalities), performance metrics across studies were not directly comparable.

    ‎
    Figure 4. Evidence gap map illustrating the frequency of machine learning (ML) methods applied to medication reconciliation tasks by publication year. BPMH: best possible medication history; CNN: convolutional neural network; CRF: conditional random field; RNN: recurrent neural network; SVM: support vector machine.

    Among the studies that used free-text data sources (72/94, 76.6%), there was variation in the granularity of information extraction. Most of these studies (57/72, 79.2%) used ML-based named entity recognition techniques to extract a variety of medication-related elements, such as dosage, route of administration, frequency, and indication [30,41,43-53,55,57,58,60,61,63,67-71,73-75,78-81,85,91-95,99,100,103,104,106-109,113,115-119,121,123,125-128]. A smaller proportion of these studies (28/72, 38.9%) extended their approach to include relation extraction, identifying associations between these elements and specific medications (eg, linking a particular dosage to a specific medication) [30,41,43-45,47,50-52,63,73,78,79,81,87,91,93,99,100,106,107,115,116,123,125-128]. An even more limited subset (9/72, 12.5%) focused on identifying medication changes (eg, discontinuation or dosage adjustment) and the context of such changes (eg, the temporality of the change and the person responsible for the change) [46,47,57,86,103,104,106,118,121]. These studies typically assessed performance using standard evaluation metrics for natural language processing information extraction tasks, including precision, recall, and F1-score.

    Studies that used image data (21/94, 22.3%) were primarily focused on medication identification. These studies frequently analyzed images of pills [42,77,82,96,97,102,112,114,120,122,124,129,130] or blister packs [66] to identify medications or generate a ranked list of possible medications. Others applied optical character recognition to pill bottle images or handwritten prescriptions and used natural language processing methods to extract medication information [59,76,101]. Most pill identification studies (10/13, 76.9%) used single-pill images, although a few (3/13, 23.1%) addressed multi-pill image identification [77,96,97]. No studies using pill images explicitly reported using patient-generated images, although many (8/13, 61.5%) incorporated intentional variations in lighting, zooming, and background, among other aspects, to simulate photos taken in real-world conditions [77,82,97,112,120,122,124,130]. The image-based studies often used more general accuracy metrics.

    Data Sources and Collection

    Most included studies (67/94, 71.3%) used clinical notes in the EHR [30,41,43-54,56,57,60-63,65,67,68,70,71,73-75,78-81,83-89,91,93-95,98-100,103-106,109-111,113,115-119,121,123,125-128], and a significant proportion of studies (21/94, 22.3%) used images [40,42,59,64,66,72,76,77,82,90,96,97,101,102,112,114,120,122,124,129,130]. In most studies using images, the images featured pills (13/21, 61.9%), although 38.1% (8/21) of the studies used images of other medication-related items (eg, medication bottles and handwritten prescriptions) [40,59,64,66,72,76,90,101]. A few studies used other sources, such as pharmacy records (3/94, 3.2%) [92,131,132], patient-physician conversation transcripts (2/94, 2.1%) [58,107], structured prescription lists in the EHR (1/94, 1.1%) [80], user-generated structured data (1/94, 1.1%) [84], user-generated notes (1/94, 1.1%) [108], or a medical internet forum (1/94, 1.1%) [48].

    Although many studies (33/94, 35.1%) used author-collected data [40,42,56,58,59,61,64-66,80,83,84,90,96-99,101,107,108,111-114,124,129,131,132], most studies used publicly available data from various benchmarking shared tasks, including i2b2 2009 (19/94, 20.2%) [48,53-55,61-63,74,79,88,89,93,95,100,110,115-117,119], n2c2 2018 (16/94, 17%) [30,41,45,49,52,71,73,85,87,91,93-95,123,127,128], n2c2 2022 (12/94, 12.8%) [46,47,57,60,61,75,103-106,118,121], MADE 1.0 2018 (7/94, 7.4%) [43,44,50,51,81,125,126], and the National Institutes of Health National Library of Medicine Pill Image Recognition (4/94, 4.3%) datasets [77,120,122,130]. The i2b2 2009 and n2c2 2018 datasets consisted solely of discharge summaries [100,134], whereas the n2c2 2022 and MADE 1.0 datasets were not restricted to a single note type [135,136]. The n2c2 2018 and 2022 datasets were annotated by at least 2 independent expert annotators per note. While MADE 1.0 also used 2 annotators, they were not independent. In contrast, the i2b2 2009 dataset was released without annotations, requiring researchers to perform their own labeling. The National Library of Medicine Pill Image Recognition dataset includes both standardized, high-quality pill images and more variable consumer-quality images to reflect real-world user-generated image conditions [137].

    Research Maturity

    Most of the included studies (93/94, 98.9%) were limited to model development, with only 1.1% (1/94) of the studies advancing to a more mature stage of implementation [84].


    Principal Findings

    Overview

    This scoping review aimed to characterize how researchers have used AI to facilitate MedRec tasks and subtasks. Notably, all the included studies (94/94, 100%) addressed subtasks related to the creation of a BPMH, but only a handful (2/94, 2.1%) developed models or tools that facilitated discrepancy identification, and none facilitated discrepancy resolution. While there is growing interest in applying AI to support MedRec tasks, most research remains focused on demonstrating technical feasibility rather than evaluating clinical integration or real-world impact. As such, efforts to automate MedRec tasks using AI are still preliminary, with current work largely focused on a limited set of subtasks, whereas other subtasks remain unexamined.

    Creation of a BPMH

    The creation of a BPMH is a critical component of MedRec. One study of inpatient MedRec found that most unintentional medication discrepancies that had the potential to cause harm were the result of medication history errors [138], highlighting the importance of this task. In addition, BPMH creation can be time-consuming [139], emphasizing the value of automating aspects of this process.

    Although all studies included in this review (94/94, 100%) developed models or tools that focused on subtasks of BPMH creation—such as medication extraction or verification—only 1.1% (1/94) developed a method that could potentially be applied to automate BPMH creation in an end-to-end manner [80]. This gap suggests that the end-to-end automation of MedRec tasks is limited not only by the state of AI but also by infrastructural barriers such as limitations to the quick exchange of medication data between EHRs and pharmacies or data incompleteness [140]. Previous work examining barriers to effective MedRec has identified the inability to access reliable medication information as a recurring challenge as users often lack access to records from external settings and the information present in EHRs may be incomplete or outdated [141-143]. These challenges may explain why so few studies (1/94, 1.1%) had progressed beyond model development to more advanced stages of implementation. Given that many models target information acquisition tasks, future research should examine what information is available to clinicians at the point of care, how these tools can augment that information, and the extent to which their use reduces cognitive and operational burden. A well-integrated medication repository that consolidates comprehensive drug histories across care settings could mitigate infrastructural challenges by improving data completeness and interoperability [144]. While such systems are not yet widely adopted or fully integrated across care settings, their expansion and improvement represent an important opportunity for future MedRec work. Efforts to update and make efficient use of health care infrastructure are key to facilitating the research, development, and implementation of tools that can scale the automation of MedRec.

    The BPMH subtasks that the included studies addressed were relatively broad in scope (ie, collecting medication information from various sources and verifying the accuracy of this information). While BPMH creation is central to MedRec, its subtasks overlap functionally with other medication-related processes. This functional overlap highlights the potential for cross-task model reuse and suggests that automation efforts need not be strictly siloed within this first MedRec stage. Many of these models have broader applications beyond MedRec, including pharmacovigilance, adherence monitoring, and medication review (ie, evaluation and optimization of a patient’s medications [6]). More broadly, automating the creation of a complete medication list may enable downstream analytical use cases such as risk stratification (eg, age-related medication risks and fall risk from sedating medications), detection of potentially inappropriate prescribing or polypharmacy, and identification of deprescribing opportunities [145-147]. This broader relevance may partially explain the sustained research interest in developing these methods, particularly as public benchmarking initiatives such as the n2c2 2018 shared task on ADE and medication extraction [134] have incentivized the development of tools that support medication data processing across use cases.

    The extent to which the included studies used data sources that reflect actual medication use is unclear, representing a key limitation in current BPMH automation research. The prevalence of patient nonadherence reported in the literature is often high [148], highlighting the importance of using sources of medication information that reflect actual use when developing a BPMH. The n2c2 2022 shared task on contextualized medication event extraction included identification of the “actor” responsible for medication changes (including patient-initiated changes) [135], suggesting that clinical notes may contain discernible indicators of actual use. Other data sources that may provide insights into actual medication use are patient-physician and patient-nurse conversation transcripts. As with clinical notes, such tools would require contextualized approaches to distinguish between mentions of prescribed regimens and actual use. Similarly, although pill image–based models could provide a closer approximation of actual use than health care professional-documented sources, studies have yet to validate them on patient-submitted images. Future research should focus on incorporating data sources that reflect actual medication use (eg, patient self-reported data) to improve the accuracy and completeness of the BPMH.

    Some tools developed may also have relevance in virtual or remote care contexts. For instance, image-based methods for medication identification could be integrated into virtual care video visits to facilitate the creation of a BPMH (eg, by acting as an additional method to verify the medications taken by patients). This functionality could serve as a supplemental verification method, which may be useful for patients with communication barriers or low health literacy or managing complex medication regimens. As health care systems continue to expand telehealth offerings, these tools may become increasingly relevant for extending MedRec capabilities outside of traditional clinical environments.

    Identification of Discrepancies

    Beyond gathering medication data to create a BPMH, later MedRec tasks often require tailored approaches to identify and resolve discrepancies. This review found one study that compared medications from clinical notes and discharge prescriptions and another that flagged discrepancies between AI-extracted and patient-entered medication lists. Only one of these studies incorporated patient-entered data as a source of medication information, reflecting actual medication use [84]. The other study that supported this task compared structured discharge prescription lists and medications extracted from free-text clinical notes [80], although it is unclear whether such notes reflect prescribed regimens, actual use, or both. The limited adoption of AI for this task may be partially related to the existence of non–AI-based electronic MedRec tools that facilitate discrepancy identification. For example, the Twinlist MedRec tool uses rule-based processing to detect discrepancies, using visual elements such as spatial grouping and highlighting to guide users in addressing these discrepancies [22]. Other tools have adopted similar strategies [149,150]. However, AI models hold promise for automating information analysis more comprehensively by enabling more robust discrepancy identification (eg, in cases in which medication information is not standardized in structure) or handling more complex medication regimens (eg, dose tapering). Given that existing tools for discrepancy identification may already be addressed in some settings, AI and health system researchers should engage with clinicians to identify what MedRec tasks or subtasks would benefit most from automation, ensuring that the tools meaningfully complement clinical workflows.

    Resolution of Discrepancies

    Discrepancy resolution represents a significant opportunity for AI innovation in MedRec as it requires interpreting and integrating contextual information that current tools do not address. Inconsistent resolution practices may contribute to the lack of clear safety benefits from MedRec, highlighting the need for standardized approaches to discrepancy resolution [151]. This process often involves determining whether a discrepancy is intentional or unintentional, a subtask that may require additional details such as the timing of a medication change. The recent n2c2 2022 shared task on contextualized medication event extraction suggests that these considerations are receiving increasing attention [135]. For example, in addition to helping discern actual medication use from clinical notes, automated contextualized extraction could facilitate the identification of health care professional-initiated medication changes as well as the nature of such changes (eg, stopping a medication), which may aid decision-making in light of discrepancies. Still, curating datasets to train models for discrepancy resolution may be particularly challenging, requiring granular annotations of clinical intent that are rarely captured in real-world records. Furthermore, other organizational barriers to MedRec, such as ambiguity in roles and lack of standardized workflows [152,153], may be less amenable to technological solutions, emphasizing the importance of evaluating tools in real-world contexts where such factors are relevant.

    The automation of decision selection in MedRec requires improved data collection practices and focused model development. To mitigate this gap without overburdening clinicians, one potential strategy is to focus structured data entry efforts on high-risk or high-variability medications where the clinical stakes are greatest, such as high-alert medications [140]. For instance, given the risks of severe hemorrhage associated with improper use of antithrombic medications [154], this medication group could be a key target for AI-assisted MedRec. Focusing on high-alert medication groups such as these would allow for more targeted model training while preserving overall workflow efficiency. Moreover, to automate this MedRec task effectively, future models must be capable of analyzing multiple data types—including clinical notes, structured orders, medication timelines, and even health care professional communication patterns—to infer intent and recommend appropriate actions. This multimodal challenge, which extends beyond collecting the correct information, remains an underexplored but essential area for advancing toward fully integrated, AI-assisted MedRec workflows.

    Limitations

    There are limitations to this review. While the conceptual framework of MedRec tasks that we used provided a useful analytic structure, MedRec workflows vary across institutions. As such, the tasks and subtasks outlined in this review may not be universally applicable. Furthermore, although we framed the creation of a BPMH as the “information acquisition” step of MedRec, this process itself draws on medication information from multiple sources (eg, patient interview, pharmacy dispensing data, and EHR prescription lists). Hence, creating a BPMH also requires elements of information analysis and decision selection steps (eg, when sources used to create a BPMH are in conflict) such that the models and tools identified in this review are not necessarily restricted in their utility to just one MedRec task. In light of existing reviews, we excluded studies primarily focused on adherence, adverse drug reactions, side effects, or allergies; however, these aspects of medication use are relevant to comprehensive MedRec efforts. We direct readers to the work by Bohlmann et al [37] and Syrowatka et al [31] for syntheses of the literature on AI or ML for adherence and ADE applications, which include studies relevant to the creation of a BPMH.

    Finally, our search did not capture literature published since mid-2024, and therefore, our findings would not reflect more recent approaches. There may also be some commercial MedRec systems that use AI but that have not been evaluated in a research context and that, therefore, our search would not have identified.

    Conclusions

    AI holds potential for supporting MedRec by reducing manual workloads and improving efficiency. This work is the first to provide a comprehensive overview of the use of AI for MedRec, using an information processing framework to evaluate the degree of automation achieved across MedRec tasks. While previous reviews have explored the use of electronic tools for MedRec, none have focused on AI applications specifically. Our review highlights which MedRec tasks have received the most AI research effort and where gaps remain, providing valuable direction for future work that aims to advance the automation of MedRec. Current AI research in MedRec remains concentrated on a limited set of subtasks, primarily at the model development stage. To fully harness the potential of AI in MedRec, researchers and health care systems should approach implementation critically, emphasizing iterative evaluation and refinement. The aim is augmentation, not replacement. Rather than a simple transfer of responsibilities from human to machine, automation should be viewed as a dynamic process featuring ongoing user and AI tool interactions, as well as ongoing verification and refinement [155-157]. Researchers should balance refining AI-based models with the prevention of medication-related harms. Future research should prioritize developing models to automate discrepancy resolution, developing and evaluating AI-based MedRec tools, and addressing infrastructural barriers to broader implementation. Health systems need usable, well-integrated tools, as well as translational studies to assess their real-world effectiveness and safety. These tools will operate within complex sociotechnical systems, where success depends not only on model performance but also on usability, interoperability, and trust.

    Funding

    This study was funded through an award from the Population Health Research Institute at Hamilton Health Sciences. The funders had no role in the collection, analysis, and interpretation of the data; writing of this paper; or decision to submit it for publication.

    Authors' Contributions

    Conceptualization: JPTB, MPB, JR, PJD, MM, JP

    Formal analysis: JPTB

    Funding acquisition: PJD, MM

    Investigation: JPTB, MPB, BW, PN

    Methodology: JPTB, MPB, JR, JP

    Project administration: JPTB, MPB

    Supervision: PJD, MM, JP

    Writing—original draft: JPTB

    Writing—review and editing: MPB, JR, BW, PN, PJD, MM, JP

    Conflicts of Interest

    None declared.

    Multimedia Appendix 1

    Full search strategies.

    DOCX File, 27 KB

    Multimedia Appendix 2

    Extraction table.

    XLSX File, 317 KB

    Checklist 1

    PRISMA-ScR checklist.

    PDF File, 743 KB

    Checklist 2

    PRISMA-S checklist.

    PDF File, 140 KB

    Checklist 3

    PRISMA 2020 for Abstracts checklist.

    PDF File, 194 KB
    1. Hakkarainen KM, Hedna K, Petzold M, Hägg S. Percentage of patients with preventable adverse drug reactions and preventability of adverse drug reactions--a meta-analysis. PLoS One. 2012;7(3):e33236. [CrossRef] [Medline]
    2. Insani WN, Whittlesea C, Alwafi H, Man KK, Chapman S, Wei L. Prevalence of adverse drug reactions in the primary care setting: a systematic review and meta-analysis. PLoS One. 2021;16(5):e0252161. [CrossRef] [Medline]
    3. Laatikainen O, Miettunen J, Sneck S, Lehtiniemi H, Tenhunen O, Turpeinen M. The prevalence of medication-related adverse events in inpatients-a systematic review and meta-analysis. Eur J Clin Pharmacol. Dec 2017;73(12):1539-1549. [CrossRef] [Medline]
    4. Elliott RA, Camacho E, Jankovic D, Sculpher MJ, Faria R. Economic analysis of the prevalence and clinical and economic burden of medication error in England. BMJ Qual Saf. Feb 2021;30(2):96-105. [CrossRef] [Medline]
    5. Medication without harm: WHO global patient safety challenge. World Health Organization. 2017. URL: https://iris.who.int/server/api/core/bitstreams/15520c4f-89d1-4a8f-9bab-6b771aa30acb/content [Accessed 2025-03-08]
    6. Penm J, Vaillancourt R, Pouliot A. Defining and identifying concepts of medication reconciliation: an international pharmacy perspective. Res Social Adm Pharm. Jun 2019;15(6):632-640. [CrossRef] [Medline]
    7. Coleman EA, Smith JD, Raha D, Min SJ. Posthospital medication discrepancies: prevalence and contributing factors. Arch Intern Med. Sep 12, 2005;165(16):1842-1847. [CrossRef] [Medline]
    8. Tully AP, Hammond DA, Li C, Jarrell AS, Kruer RM. Evaluation of medication errors at the transition of care from an ICU to non-ICU location. Crit Care Med. Apr 2019;47(4):543-549. [CrossRef] [Medline]
    9. Alqenae FA, Steinke D, Keers RN. Prevalence and nature of medication errors and medication-related harm following discharge from hospital to community settings: a systematic review. Drug Saf. Jun 2020;43(6):517-537. [CrossRef] [Medline]
    10. Quélennec B, Beretz L, Paya D, et al. Potential clinical impact of medication discrepancies at hospital admission. Eur J Intern Med. Sep 2013;24(6):530-535. [CrossRef] [Medline]
    11. McNab D, Bowie P, Ross A, MacWalter G, Ryan M, Morrison J. Systematic review and meta-analysis of the effectiveness of pharmacist-led medication reconciliation in the community after hospital discharge. BMJ Qual Saf. Apr 2018;27(4):308-320. [CrossRef] [Medline]
    12. Required Organizational Practices 2021 Handbook. Accreditation Canada; 2021.
    13. National Patient Safety Goals®: effective January 2025 for the hospital program. The Joint Commission. 2024. URL: https:/​/digitalassets.​jointcommission.org/​api/​public/​content/​9be383450fc941df806b76c5fbdd9ae6?v=3c600c3a [Accessed 2026-04-29]
    14. Medicines optimisation. National Institute for Health and Care Excellence. 2016. URL: https://www.nice.org.uk/guidance/qs120/resources/medicines-optimisation-pdf-75545351857861 [Accessed 2026-04-29]
    15. National safety and quality health service standards. Australian Commission on Safety and Quality in Health Care. 2012. URL: https:/​/www.​spkc.gov.lv/​sites/​spkc/​files/​pacientu_drosiba/​standarti/​australia_nsqhsstandardssept2012.​pdf [Accessed 2026-04-29]
    16. Mekonnen AB, McLachlan AJ, Brien JA. Effectiveness of pharmacist-led medication reconciliation programmes on clinical outcomes at hospital transitions: a systematic review and meta-analysis. BMJ Open. Feb 23, 2016;6(2):e010003. [CrossRef] [Medline]
    17. Cheema E, Alhomoud FK, Kinsara AS, et al. The impact of pharmacists-led medicines reconciliation on healthcare outcomes in secondary care: a systematic review and meta-analysis of randomized controlled trials. PLoS One. 2018;13(3):e0193510. [CrossRef] [Medline]
    18. Redmond P, Grimes TC, McDonnell R, Boland F, Hughes C, Fahey T. Impact of medication reconciliation for improving transitions of care. Cochrane Database Syst Rev. Aug 23, 2018;8(8):CD010791. [CrossRef] [Medline]
    19. Etchells E, Fernandes O. Medication reconciliation: ineffective or hard to implement? BMJ Qual Saf. Dec 2018;27(12):947-949. [CrossRef] [Medline]
    20. Pevnick JM, Shane R, Schnipper JL. The problem with medication reconciliation. BMJ Qual Saf. Sep 2016;25(9):726-730. [CrossRef] [Medline]
    21. Meguerditchian AN, Krotneva S, Reidel K, Huang A, Tamblyn R. Medication reconciliation at admission and discharge: a time and motion study. BMC Health Serv Res. Nov 21, 2013;13:485. [CrossRef] [Medline]
    22. Plaisant C, Chao T, Wu J, et al. Twinlist: novel user interface designs for medication reconciliation. AMIA Annu Symp Proc. 2013;2013:1150-1159. [Medline]
    23. Tamblyn R, Winslade N, Lee TC, et al. Improving patient safety and efficiency of medication reconciliation through the development and adoption of a computer-assisted tool with automated electronic integration of population-based community drug data: the RightRx project. J Am Med Inform Assoc. May 1, 2018;25(5):482-495. [CrossRef] [Medline]
    24. Gadallah A, McGinnis B, Nguyen B, Olson J. Assessing the impact of virtual medication history technicians on medication reconciliation discrepancies. Int J Clin Pharm. Oct 2021;43(5):1404-1411. [CrossRef] [Medline]
    25. Schubert CC, Penney LS, Schwartzkopf AL, et al. Expanding access to comprehensive geriatric evaluation via telehealth: development of hybrid-virtual home visits. J Gen Intern Med. Feb 2024;39(Suppl 1):36-43. [CrossRef] [Medline]
    26. Heyworth L, Paquin AM, Clark J, et al. Engaging patients in medication reconciliation via a patient portal following hospital discharge. J Am Med Inform Assoc. Feb 2014;21(e1):e157-e162. [CrossRef] [Medline]
    27. Zayas-Cabán T, Okubo TH, Posnack S. Priorities to accelerate workflow automation in health care. J Am Med Inform Assoc. 2022;30(1):195-201. [CrossRef]
    28. Marien S, Krug B, Spinewine A. Electronic tools to support medication reconciliation: a systematic review. J Am Med Inform Assoc. Jan 2017;24(1):227-240. [CrossRef] [Medline]
    29. Ciudad-Gutiérrez P, Del Valle-Moreno P, Lora-Escobar SJ, Guisado-Gil AB, Alfaro-Lara ER. Electronic medication reconciliation tools aimed at healthcare professionals to support medication reconciliation: a systematic review. J Med Syst. Dec 6, 2023;48(1):2. [CrossRef] [Medline]
    30. Christopoulou F, Tran TT, Sahu SK, Miwa M, Ananiadou S. Adverse drug events and medication relation extraction in electronic health records with ensemble deep learning methods. J Am Med Inform Assoc. Jan 1, 2020;27(1):39-46. [CrossRef] [Medline]
    31. Syrowatka A, Song W, Amato MG, et al. Key use cases for artificial intelligence to reduce the frequency of adverse drug events: a scoping review. Lancet Digit Health. Feb 2022;4(2):e137-e148. [CrossRef] [Medline]
    32. Yao X, Rao A, Padman R. Analytical approaches for medication management for patient safety: a scoping review. NPJ Health Syst. 2025;2:46. [CrossRef]
    33. Ong JC, Chen MH, Ng N, et al. A scoping review on generative AI and large language models in mitigating medication related harm. NPJ Digit Med. Mar 28, 2025;8:182. [CrossRef]
    34. Peters MD, Godfrey C, McInerney P, Munn Z, Tricco AC, Khalil H. Scoping reviews. In: Aromataris E, Lockwood C, Porritt K, Pilla B, Jordan Z, editors. JBI Manual for Evidence Synthesis. JBI; 2024. [CrossRef]
    35. Tricco AC, Lillie E, Zarin W, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med. Oct 2, 2018;169(7):467-473. [CrossRef] [Medline]
    36. Tabja Bortesi J, Ranisau J. Automation in AI-based medication reconciliation tools: a scoping review protocol. OSF. Preprint posted online on Jun 10, 2024. [CrossRef]
    37. Bohlmann A, Mostafa J, Kumar M. Machine learning and medication adherence: scoping review. JMIRx Med. Nov 24, 2021;2(4):e26993. [CrossRef] [Medline]
    38. Haddaway NR, Grainger MJ, Gray CT. Citationchaser: a tool for transparent and efficient forward and backward citation chasing in systematic searching. Res Synth Methods. Jul 2022;13(4):533-545. [CrossRef] [Medline]
    39. Rethlefsen ML, Kirtley S, Waffenschmidt S, et al. PRISMA-S: an extension to the PRISMA statement for reporting literature searches in systematic reviews. Syst Rev. Jan 26, 2021;10(1):39. [CrossRef] [Medline]
    40. Achkar R, Ghayad K, Haidar R, Saleh S, Al Hajj R. Medical handwritten prescription recognition using CRNN. Presented at: 2019 International Conference on Computer, Information and Telecommunication Systems (CITS); Aug 28-31, 2019. [CrossRef]
    41. Alfattni G, Belousov M, Peek N, Nenadic G. Extracting drug names and associated attributes from discharge summaries: text mining study. JMIR Med Inform. May 5, 2021;9(5):e24678. [CrossRef] [Medline]
    42. Ashraf AR, Somogyi-Végh A, Merczel S, Gyimesi N, Fittler A. Leveraging code-free deep learning for pill recognition in clinical settings: a multicenter, real-world study of performance across multiple platforms. Artif Intell Med. Apr 2024;150:102844. [CrossRef] [Medline]
    43. Chapman AB, Peterson KS, Alba PR, DuVall SL, Patterson OV. Hybrid system for adverse drug event detection. In: Proceedings of the 1st International Workshop on Medication and Adverse Drug Event Detection. Vol 90. PMLR; 2018:16-24. URL: https://proceedings.mlr.press/v90/chapman18a.html [Accessed 2025-04-21]
    44. Chapman AB, Peterson KS, Alba PR, DuVall SL, Patterson OV. Detecting adverse drug events with rapidly trained classification models. Drug Saf. Jan 2019;42(1):147-156. [CrossRef] [Medline]
    45. Chen L, Gu Y, Ji X, et al. Extracting medications and associated adverse drug events using a natural language processing system combining knowledge base and deep learning. J Am Med Inform Assoc. Jan 1, 2020;27(1):56-64. [CrossRef] [Medline]
    46. Chen TY, Aquino JA, Chiu YW, Yeh WC, Chang YC. Understanding the clinical context of medication change events in clinical narratives using pre-trained clinical language models. Presented at: ICMHI ’23: Proceedings of the 2023 7th International Conference on Medical and Health Informatics; May 12-14, 2023. [CrossRef]
    47. Chen A, Yu Z, Yang X, Guo Y, Bian J, Wu Y. Contextualized medication information extraction using transformer-based deep learning architectures. J Biomed Inform. Jun 2023;142:104370. [CrossRef] [Medline]
    48. Dai X, Karimi S, Paris C. Medication and adverse event extraction from noisy text. Presented at: Proceedings of the Australasian Language Technology Association Workshop 2017; Dec 6-8, 2017. URL: https://aclanthology.org/U17-1009/ [Accessed 2026-04-30]
    49. Dai HJ, Su CH, Wu CS. Adverse drug event and medication extraction in electronic health records via a cascading architecture with different sequence labeling models and word embeddings. J Am Med Inform Assoc. Jan 1, 2020;27(1):47-55. [CrossRef] [Medline]
    50. Dandala B, Joopudi V, Devarakonda M. IBM research system at MADE 2018: detecting adverse drug events from electronic health records. In: Proceedings of the 1st International Workshop on Medication and Adverse Drug Event Detection. Vol 90. PMLR; 2018:39-47. URL: https://proceedings.mlr.press/v90/dandala18a.html [Accessed 2025-04-21]
    51. Dandala B, Joopudi V, Devarakonda M. Adverse drug events detection in clinical notes by jointly modeling entities and relations using neural networks. Drug Saf. Jan 2019;42(1):135-146. [CrossRef] [Medline]
    52. Dandala B, Joopudi V, Tsou CH, Liang JJ, Suryanarayanan P. Extraction of information related to drug safety surveillance from electronic health record notes: joint modeling of entities and relations using knowledge-aware neural attentive models. JMIR Med Inform. Jul 10, 2020;8(7):e18417. [CrossRef] [Medline]
    53. Doan S, Xu H. Recognizing medication related entities in hospital discharge summaries using support vector machine. Proc Int Conf Comput Ling. Aug 2010;2010:259-266. [Medline]
    54. Doan S, Bastarache L, Klimkowski S, Denny JC, Xu H. Integrating existing natural language processing tools for medication extraction from discharge summaries. J Am Med Inform Assoc. 2010;17(5):528-531. [CrossRef] [Medline]
    55. Doan S, Collier N, Xu H, Pham HD, Tu MP. Recognition of medication information from discharge summaries using ensembles of classifiers. BMC Med Inform Decis Mak. May 7, 2012;12:36. [CrossRef] [Medline]
    56. Fan Y, Zhang R. Using natural language processing methods to classify use status of dietary supplements in clinical notes. BMC Med Inform Decis Mak. Jul 23, 2018;18(Suppl 2):51. [CrossRef] [Medline]
    57. Gan Q, Hu M, Peterson KS, et al. A deep learning approach for medication disposition and corresponding attributes extraction. J Biomed Inform. Jul 2023;143:104391. [CrossRef] [Medline]
    58. Ganoe CH, Wu W, Barr PJ, et al. Natural language processing for automated annotation of medication mentions in primary care visit conversations. JAMIA Open. 2021;4(3):ooab071. [CrossRef] [Medline]
    59. Gromova K, Elangovan V. Automatic extraction of medication information from cylindrically distorted pill bottle labels. Mach Learn Knowl Extr. 2022;4(4):852-864. [CrossRef]
    60. Guo Y, Ge Y, Sarker A. Detection of medication mentions and medication change events in clinical notes using transformer-based models. Stud Health Technol Inform. Jan 25, 2024;310:685-689. [CrossRef] [Medline]
    61. Guzman B, Metzger I, Aphinyanaphongs Y, Grover H. Assessment of Amazon Comprehend Medical: medication information extraction. arXiv. Preprint posted online on Feb 2, 2020. [CrossRef]
    62. Halgrim S, Xia F, Solti I, Cadag E, Uzuner Ö. Extracting medication information from discharge summaries. In: Dalianis H, Hassel M, Nilsson G, editors. Proceedings of the NAACL HLT 2010 Second Louhi Workshop on Text and Data Mining of Health Documents Association for Computational Linguistics. Association for Computational Linguistics; 2010:61-67.
    63. Halgrim SR, Xia F, Solti I, Cadag E, Uzuner Ö. A cascade of classifiers for extracting medication information from discharge summaries. J Biomed Semant. 2011;2(Suppl 3):S2. [CrossRef]
    64. Hassan E, Tarek H, Hazem M, Bahnacy S, Shaheen L, Elashmwai WH. Medical prescription recognition using machine learning. Presented at: 2021 IEEE 11th Annual Computing and Communication Workshop and Conference (CCWC); Jan 27-20, 2021. [CrossRef]
    65. Iglesias JE, Rocks K, Jahanshad N, Frias-Martinez E, Andrada LP, Bui AA. Tracking medication information across medical records. AMIA Annu Symp Proc. Nov 14, 2009;2009:266-270. [Medline]
    66. Itagi A, Sil R, Mohapatra S, et al. Medicine strip identification using 2-D cepstral feature extraction and multiclass classification methods. arXiv. Preprint posted online on Feb 3, 2020. [CrossRef]
    67. Jagannatha AN, Yu H. Structured prediction models for RNN based sequence labeling in clinical text. Proc Conf Empir Methods Nat Lang Process. Nov 2016;2016:856-865. [CrossRef] [Medline]
    68. Jagannatha AN, Yu H. Bidirectional RNN for medical event detection in electronic health records. Proc Conf. Jun 2016;2016:473-482. [CrossRef] [Medline]
    69. Jagannathan V, Mullett CJ, Arbogast JG, et al. Assessment of commercial NLP engines for medication information extraction from dictated clinical notes. Int J Med Inform. Apr 2009;78(4):284-291. [CrossRef] [Medline]
    70. Jouffroy J, Feldman SF, Lerner I, Rance B, Burgun A, Neuraz A. Hybrid deep learning for medication-related information extraction from clinical texts in French: MedExt algorithm development study. JMIR Med Inform. Mar 16, 2021;9(3):e17934. [CrossRef] [Medline]
    71. Ju M, Nguyen NT, Miwa M, Ananiadou S. An ensemble of neural models for nested adverse drug events and medication extraction with subwords. J Am Med Inform Assoc. Jan 1, 2020;27(1):22-30. [CrossRef] [Medline]
    72. Kamalanaban E, Gopinath M, Premkumar S. Medicine box: doctor’s prescription recognition using deep machine learning. Int J Eng Technol. 2018;7(3):114. [CrossRef]
    73. Kim Y, Meystre SM. Ensemble method-based extraction of medication and related information from clinical texts. J Am Med Inform Assoc. Jan 1, 2020;27(1):31-38. [CrossRef] [Medline]
    74. Kocabiyikoglu AC, Babouchkine JM, Portet F, Qader R. Neural medication extraction: a comparison of recent models in supervised and semi-supervised learning settings. Presented at: 2021 IEEE 9th International Conference on Healthcare Informatics (ICHI); Aug 9-12, 2021. [CrossRef]
    75. Koontz J, Oronoz M, Pérez A. Evaluating data augmentation for medication identification in clinical notes. In: Mitkov R, Angelova G, editors. Proceedings of the 14th International Conference on Recent Advances in Natural Language Processing. INCOMA:578-585. [CrossRef]
    76. Kulathunga D, Muthukumarana C, Pasan U, Hemachandra C, Tissera M, De Silva H. PatientCare: patient assistive tool with automatic hand-written prescription reader. Presented at: 2nd International Conference on Advancements in Computing (ICAC); Dec 10-11, 2020. [CrossRef]
    77. Larios Delgado N, Usuyama N, Hall AK, et al. Fast and accurate medication identification. NPJ Digit Med. 2019;2:10. [CrossRef] [Medline]
    78. Lerner I, Jouffroy J, Burgun A, Neuraz A. Learning the grammar of drug prescription: recurrent neural network grammars for medication information extraction in clinical texts. arXiv. Preprint posted online on Apr 22, 2022. [CrossRef]
    79. Li Z, Liu F, Antieau L, Cao Y, Yu H. Lancet: a high precision medication event extraction system for clinical text. J Am Med Inform Assoc. 2010;17(5):563-567. [CrossRef] [Medline]
    80. Li Q, Spooner SA, Kaiser M, et al. An end-to-end hybrid algorithm for automated medication discrepancy detection. BMC Med Inform Decis Mak. May 6, 2015;15:37. [CrossRef] [Medline]
    81. Li F, Liu W, Yu H. Extraction of information related to adverse drug events from electronic health record notes: design of an end-to-end model based on deep learning. JMIR Med Inform. Nov 26, 2018;6(4):e12159. [CrossRef] [Medline]
    82. Ling S, Pastor A, Li J, et al. Few-shot pill recognition. Presented at: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR); Jun 13-19, 2020. [CrossRef]
    83. Liu F, Pradhan R, Druhl E, et al. Learning to detect and understand drug discontinuation events from clinical narratives. J Am Med Inform Assoc. Oct 1, 2019;26(10):943-951. [CrossRef] [Medline]
    84. Long J, Yuan MJ, Poonawala R. An observational study to evaluate the usability and intent to adopt an artificial intelligence-powered medication reconciliation tool. Interact J Med Res. May 16, 2016;5(2):e14. [CrossRef] [Medline]
    85. Mahajan D, Liang JJ, Tsou CH. Extracting daily dosage from medication instructions in EHRs: an automated approach and lessons learned. arXiv. Preprint posted online on May 21, 2020. [CrossRef]
    86. Mahajan D, Liang JJ, Tsou CH. Toward understanding clinical context of medication change events in clinical narratives. AMIA Annu Symp Proc. 2022;2021:833-842. [Medline]
    87. Mahendran D, McInnes BT. Extracting adverse drug events from clinical notes. AMIA Jt Summits Transl Sci Proc. 2021;2021:420-429. [Medline]
    88. Meystre SM, Thibault J, Shen S, Hurdle JF, South BR. Textractor: a hybrid system for medications and reason for their prescription extraction from clinical text documents. J Am Med Inform Assoc. 2010;17(5):559-562. [CrossRef] [Medline]
    89. Meystre SM, Thibault J, Shen S, Hurdle JF, South BR. Automatically detecting medications and the reason for their prescription in clinical narrative text documents. Stud Health Technol Inform. 2010;160(Pt 2):944-948. [Medline]
    90. Miao J, Huang Y, Wang Z, Wu Z, Lv J. Image recognition of traditional Chinese medicine based on deep learning. Front Bioeng Biotechnol. 2023;11:1199803. [CrossRef] [Medline]
    91. Miller T, Geva A, Dligach D. Extracting adverse drug event information with minimal engineering. Proc Conf. Jun 2019;2019:22-27. [CrossRef] [Medline]
    92. Mishra RK, Roy S, Palla SK, Patel N, Patel M, Jos S. Hybrid approach combining deep learning and a rule based expert system for concept extraction from prescriptions. Annu Int Conf IEEE Eng Med Biol Soc. Jul 2023;2023:1-4. [CrossRef] [Medline]
    93. Modersohn L, Hahn U. Influence of context in transformer-based medication relation extraction. Stud Health Technol Inform. Jan 25, 2024;310:669-673. [CrossRef] [Medline]
    94. Narayanan S, Mannam K, Rajan SP, Rangan PV. Evaluation of transfer learning for adverse drug event (ADE) and medication entity extraction. In: Rumshisky A, Roberts K, Bethard S, Naumann T, editors. Proceedings of the 3rd Clinical Natural Language Processing Workshop. Association for Computational Linguistics; 2020:55-64. [CrossRef]
    95. Narayanan S, Mannam K, Achan P, Ramesh MV, Rangan PV, Rajan SP. A contextual multi-task neural approach to medication and adverse events identification from clinical text. J Biomed Inform. Jan 2022;125:103960. [CrossRef] [Medline]
    96. Nguyen AD, Nguyen TD, Pham HH, Nguyen TH, Nguyen PL. Image-based contextual pill recognition with medical knowledge graph assistance. arXiv. Preprint posted online on Aug 4, 2022. [CrossRef]
    97. Nguyen AD, Pham HH, Trung HT, Nguyen QV, Truong TN, Nguyen PL. High accurate and explainable multi-pill detection framework with graph neural network-assisted multimodal data fusion. PLoS One. 2023;18(9):e0291865. [CrossRef] [Medline]
    98. Pakhomov SV, Ruggieri A, Chute CG. Maximum entropy modeling for mining patient medication status from free text. Proc AMIA Symp. 2002:587-591. [Medline]
    99. Patrick J, Li M. A cascade approach to extracting medication events. Presented at: Proceedings of the Australasian Language Technology Association Workshop 2009; Dec 3-4, 2009. URL: https://aclanthology.org/U09-1014/ [Accessed 2026-04-30]
    100. Patrick J, Li M. High accuracy information extraction of medication information from clinical notes: 2009 i2b2 medication extraction challenge. J Am Med Inform Assoc. 2010;17(5):524-527. [CrossRef] [Medline]
    101. Paul S, Madan G, Mishra A, Hegde N, Kumar P, Aggarwal G. Weakly supervised information extraction from inscrutable handwritten document images. arXiv. Preprint posted online on Jun 12, 2023. [CrossRef]
    102. Pilania U, Kumar M, Singh S. Leveraging convolutional neural networks for improved drug identification. Presented at: 2023 6th International Conference on Contemporary Computing and Informatics (IC3I); Sep 14-16, 2023. [CrossRef]
    103. Ramachandran GK, Lybarger K, Liu Y, et al. Extracting medication changes in clinical narratives using pre-trained language models. J Biomed Inform. Mar 2023;139:104302. [CrossRef] [Medline]
    104. Salek Faramarzi N, Patel M, Bandarupally SH, Banerjee R. Context-aware medication event extraction from unstructured text. In: Naumann T, Ben Abacha A, Bethard S, Roberts K, Rumshisky A, editors. Proceedings of the 5th Clinical Natural Language Processing Workshop. Association for Computational Linguistics; 2023:86-95. [CrossRef]
    105. Sarker S, Qian L, Dong X. Medical data augmentation via ChatGPT: a case study on medication identification and medication event classification. arXiv. Preprint posted online on Jun 10, 2023. [CrossRef]
    106. Schäfer H, Idrissi-Yaghir A, Bewersdorff J, Frihat S, Friedrich CM, Zesch T. Medication event extraction in clinical notes: contribution of the WisPerMed team to the n2c2 2022 challenge. J Biomed Inform. Jul 2023;143:104400. [CrossRef] [Medline]
    107. Selvaraj SP, Konam S. Medication regimen extraction from medical conversations. arXiv. Preprint posted online on Dec 10, 2019. [CrossRef]
    108. Sezgin E, Hussain SA, Rust S, Huang Y. Extracting medical information from free-text and unstructured patient-generated health data using natural language processing methods: feasibility study with real-world data. JMIR Form Res. Mar 7, 2023;7:e43014. [CrossRef] [Medline]
    109. Shah-Mohammadi F, Cui W, Finkelstein J. Comparison of ACM and CLAMP for entity extraction in clinical notes. Annu Int Conf IEEE Eng Med Biol Soc. 2021;2021:1989-1992. [CrossRef]
    110. Shivade C, Cormack J, Milward D. Precise medication extraction using agile text mining. In: Velupillai S, Duneld M, Kvist M, Dalianis H, Skeppstedt M, Henriksson A, editors. Proceedings of the 5th International Workshop on Health Text Mining and Information Analysis (Louhi). Association for Computational Linguistics; 2014:75-79. [CrossRef]
    111. Sohn S, Murphy SP, Masanz JJ, Kocher JP, Savova GK. Classification of medication status change in clinical narratives. AMIA Annu Symp Proc. Nov 13, 2010;2010:762-766. [Medline]
    112. Srikamdee S, Suksawatchon U, Suksawatchon J, Werapan W. ClinicYA: an application for pill identification using deep learning and k-means clustering. Presented at: 2022 26th International Computer Science and Engineering Conference (ICSEC); Dec 21-23, 2022. [CrossRef]
    113. Stevens M, Kennedy G, Churches T. Applying and improving a publicly available medication NER pipeline in a clinical cancer EMR. Stud Health Technol Inform. Jan 25, 2024;310:679-684. [CrossRef] [Medline]
    114. Swastika W, Prilianti K, Stefanus A, et al. Preliminary study of multi convolution neural network-based model to identify pills image using classification rules. Presented at: 2019 International Seminar on Intelligent Technology and Its Applications (ISITIA); Aug 28-29, 2019. [CrossRef]
    115. Tao C, Filannino M, Uzuner Ö. Prescription extraction using CRFs and word embeddings. J Biomed Inform. Aug 2017;72:60-66. [CrossRef] [Medline]
    116. Tao C, Filannino M, Uzuner Ö. FABLE: a semi-supervised prescription information extraction system. AMIA Annu Symp Proc. 2018;2018:1534-1543. [Medline]
    117. Tikk D, Solt I. Improving textual medication extraction using combined conditional random fields and rule-based systems. J Am Med Inform Assoc. 2010;17(5):540-544. [CrossRef] [Medline]
    118. Tsujimura T, Yamada K, Ida R, Miwa M, Sasaki Y. Contextualized medication event extraction with striding NER and multi-turn QA. J Biomed Inform. Aug 2023;144:104416. [CrossRef] [Medline]
    119. Tu H, Han L, Nenadic G. Extraction of medication and temporal relation from clinical text using neural language models. Presented at: 2023 IEEE International Conference on Big Data (BigData); Dec 15-18, 2023. [CrossRef]
    120. Usuyama N, Delgado NL, Hall AK, Lundin J. ePillID dataset: a low-shot fine-grained benchmark for pill identification. Presented at: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW); Jun 14-19, 2020. [CrossRef]
    121. Vasilakes J, Georgiadis P, Nguyen NT, Miwa M, Ananiadou S. Contextualized medication event extraction with levitated markers. J Biomed Inform. May 2023;141:104347. [CrossRef] [Medline]
    122. Wang Y, Ribera J, Liu C, Yarlagadda S, Zhu F. Pill recognition using minimal labeled data. Presented at: 2017 IEEE Third International Conference on Multimedia Big Data (BigMM); Apr 19-21, 2017. [CrossRef]
    123. Wei Q, Ji Z, Li Z, et al. A study of deep learning approaches for medication and adverse drug event extraction from clinical text. J Am Med Inform Assoc. Jan 1, 2020;27(1):13-21. [CrossRef] [Medline]
    124. Wong YF, Ng HT, Leung KY, Chan KY, Chan SY, Loy CC. Development of fine-grained pill identification algorithm using deep convolutional network. J Biomed Inform. Oct 2017;74:130-136. [CrossRef] [Medline]
    125. Xu D, Yadav V, Bethard S. UArizona at the MADE1.0 NLP challenge. Proc Mach Learn Res. May 2018;90:57-65. [Medline]
    126. Yang X, Bian J, Gong Y, Hogan WR, Wu Y. MADEx: a system for detecting medications, adverse drug events, and their relations from clinical notes. Drug Saf. Jan 2019;42(1):123-133. [CrossRef] [Medline]
    127. Yang X, Bian J, Fang R, Bjarnadottir RI, Hogan WR, Wu Y. Identifying relations of medications with adverse drug events using recurrent convolutional neural networks and gradient boosting. J Am Med Inform Assoc. Jan 1, 2020;27(1):65-72. [CrossRef] [Medline]
    128. Yazdani A, Proios D, Rouhizadeh H, Teodoro D. Efficient joint learning for clinical named entity recognition and relation extraction using Fourier networks: a use case in adverse drug events. arXiv. Preprint posted online on Feb 8, 2023. [CrossRef]
    129. Yu J, Chen Z, Kamata SI, Yang J. Accurate system for automatic pill recognition using imprint information. IET Image Process. Dec 2015;9(12):1039-1047. [CrossRef]
    130. Zeng X, Cao K, Zhang M. MobileDeepPill: a small-footprint mobile deep learning system for recognizing unconstrained pill images. Presented at: MobiSys ’17: Proceedings of the 15th Annual International Conference on Mobile Systems, Applications, and Services; Jun 19-23, 2017. [CrossRef]
    131. Hasan S, Duncan GT, Neill DB, Padman R. Automatic detection of omissions in medication lists. J Am Med Inform Assoc. 2011;18(4):449-458. [CrossRef] [Medline]
    132. Hasan S, Duncan GT, Neill DB, Padman R. Towards a collaborative filtering approach to medication reconciliation. AMIA Annu Symp Proc. Nov 6, 2008;2008:288-292. [Medline]
    133. Parasuraman R, Sheridan TB, Wickens CD. A model for types and levels of human interaction with automation. IEEE Trans Syst Man Cybern A Syst Hum. May 2000;30(3):286-297. [CrossRef]
    134. Henry S, Buchan K, Filannino M, Stubbs A, Uzuner O. 2018 n2c2 shared task on adverse drug events and medication extraction in electronic health records. J Am Med Inform Assoc. Jan 1, 2020;27(1):3-12. [CrossRef] [Medline]
    135. Mahajan D, Liang JJ, Tsou CH, Uzuner Ö. Overview of the 2022 n2c2 shared task on contextualized medication event extraction in clinical notes. J Biomed Inform. Aug 2023;144:104432. [CrossRef] [Medline]
    136. Jagannatha A, Liu F, Liu W, Yu H. Overview of the first natural language processing challenge for extracting medication, indication, and adverse drug events from electronic health record notes (MADE 1.0). Drug Saf. Jan 2019;42(1):99-111. [CrossRef] [Medline]
    137. Yaniv Z, Faruque J, Howe S, et al. The National Library of Medicine pill image recognition challenge: an initial report. IEEE Appl Imag Pattern Recognit Workshop. Oct 2016;2016. [CrossRef] [Medline]
    138. Pippins JR, Gandhi TK, Hamann C, et al. Classifying and predicting errors of inpatient medication reconciliation. J Gen Intern Med. Sep 2008;23(9):1414-1422. [CrossRef] [Medline]
    139. Alix L, Bajeux E, Hubert J, et al. Medication reconciliation in hospital patients over the age of 65: how long does it take and how much does it cost? A time-motion study in an internal medicine ward. Eur J Intern Med. Mar 2020;73:100-102. [CrossRef] [Medline]
    140. Kashyap N, Jeffery S, Agresta T. From MedWreck to MedRec: a call to action to improve medication reconciliation. Appl Clin Inform. Mar 2024;15(2):230-233. [CrossRef] [Medline]
    141. Boockvar KS, Santos SL, Kushniruk A, Johnson C, Nebeker JR. Medication reconciliation: barriers and facilitators from the perspectives of resident physicians and pharmacists. J Hosp Med. 2011;6(6):329-337. [CrossRef] [Medline]
    142. Kennelty KA, Chewning B, Wise M, Kind A, Roberts T, Kreling D. Barriers and facilitators of medication reconciliation processes for recently discharged patients from community pharmacists’ perspectives. Res Social Adm Pharm. 2015;11(4):517-530. [CrossRef] [Medline]
    143. Snyder ME, Nguyen KA, Patel H, et al. Clinicians’ use of health information exchange technologies for medication reconciliation in the U.S. Department of Veterans Affairs: a qualitative analysis. BMC Health Serv Res. Oct 8, 2024;24(1):1194. [CrossRef] [Medline]
    144. Salgado-Baez E, Heidepriem R, Delucchi Danhier R, et al. Toward interoperable digital medication records on fast healthcare interoperability resources: development and technical validation of a minimal core dataset. JMIR Med Inform. May 9, 2025;13:e64099. [CrossRef] [Medline]
    145. Scott IA, Hilmer SN, Reeve E, et al. Reducing inappropriate polypharmacy: the process of deprescribing. JAMA Intern Med. May 2015;175(5):827-834. [CrossRef] [Medline]
    146. Socrates V, Wright DS, Huang T, et al. Identifying deprescribing opportunities with large language models in older adults: retrospective cohort study. JMIR Aging. Apr 11, 2025;8:e69504. [CrossRef] [Medline]
    147. Wabe N, Meulenbroeks I, Huang G, et al. Development and internal validation of a dynamic fall risk prediction and monitoring tool in aged care using routinely collected electronic health data: a landmarking approach. J Am Med Inform Assoc. Apr 19, 2024;31(5):1113-1125. [CrossRef] [Medline]
    148. Chapman SC, Chan AH. Medication nonadherence - definition, measurement, prevalence, and causes: reflecting on the past 20 years and looking forwards. Front Pharmacol. 2025;16:1465059. [CrossRef] [Medline]
    149. Marien S, Legrand D, Ramdoyal R, et al. A user-centered design and usability testing of a web-based medication reconciliation application integrated in an eHealth network. Int J Med Inform. Jun 2019;126:138-146. [CrossRef] [Medline]
    150. Gionfriddo MR, Hu Y, Maddineni B, et al. Evaluation of a web-based medication reconciliation application within a primary care setting: cluster-randomized controlled trial. JMIR Form Res. Mar 8, 2022;6(3):e33488. [CrossRef] [Medline]
    151. MacLaren R. Resolving medication reconciliation. Can J Hosp Pharm. 2022;75(3):155-156. [CrossRef] [Medline]
    152. van Sluisveld N, Zegers M, Natsch S, Wollersheim H. Medication reconciliation at hospital admission and discharge: insufficient knowledge, unclear task reallocation and lack of collaboration as major barriers to medication safety. BMC Health Serv Res. Jun 21, 2012;12:170. [CrossRef] [Medline]
    153. Gionfriddo MR, Duboski V, Middernacht A, Kern MS, Graham J, Wright EA. A mixed methods evaluation of medication reconciliation in the primary care setting. PLoS One. 2021;16(12):e0260882. [CrossRef] [Medline]
    154. High-alert medications list. ISMP Canada. URL: https://ismpcanada.ca/resource/highalertlist/ [Accessed 2026-03-26]
    155. Hopkins D, Schwanen T. Talking about automated vehicles: what do levels of automation do? Technol Soc. Feb 2021;64:101488. [CrossRef]
    156. Mosier KL, Skitka LJ. Human decision makers and automated decision aids: made for each other. In: Parasuraman R, Mouloua M, editors. Automation and Human Performance. CRC Press; 1996.
    157. Lyell D, Coiera E. Automation bias and verification complexity: a systematic review. J Am Med Inform Assoc. Mar 1, 2017;24(2):423-431. [CrossRef] [Medline]

    ‎
    ADE: adverse drug event
    AI: artificial intelligence
    BPMH: best possible medication history
    CRF: conditional random field
    EHR: electronic health record
    MedRec: medication reconciliation
    ML: machine learning
    PRISMA-S: Preferred Reporting Items for Systematic Reviews and Meta-Analyses literature search extension
    PRISMA-ScR: Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews

    Edited by Stefano Brini; submitted 29.Oct.2025; peer-reviewed by Joshua Odeniya, Kevin Fuji, Robert Marshall; final revised version received 17.Apr.2026; accepted 20.Apr.2026; published 11.May.2026.

    Copyright

    © Juan Pablo Tabja Bortesi, Maria P Becerra, Jonathan Ranisau, Bonnie Wen, Praveen Nadesan, P J Devereaux, Michael McGillion, Jeremy Petch. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 11.May.2026.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research (ISSN 1438-8871), is properly cited. The complete bibliographic information, a link to the original publication on https://www.jmir.org/, as well as this copyright and license information must be included.