Qualitative Changes in Clinical Records After Implementation of Pharmacist-Led Antimicrobial Stewardship Program: A Text Mining Analysis

Qualitative Changes in Clinical Records After Implementation of Pharmacist-Led Antimicrobial Stewardship Program: A Text Mining Analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Qualitative Changes in Clinical Records After Implementation of Pharmacist-Led Antimicrobial Stewardship Program: A Text Mining Analysis Keisuke Sawada, Shuji Kono, Ryo Inose, Yuichi Muraki This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6199166/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 23 Apr, 2025 Read the published version in Journal of Pharmaceutical Health Care and Sciences → Version 1 posted You are reading this latest preprint version Abstract Background Antimicrobial stewardship programs (ASPs) are essential for optimizing antimicrobial use, but many medium-sized hospitals lack infectious disease (ID) specialists. Ward pharmacists can contribute to ASPs, but the qualitative changes in their practice patterns after ASP implementation remains unclear. We aimed to explore the potential of text mining as a novel methodology to evaluate changes in ward pharmacist antimicrobial management practices after ASP implementation in a medium-sized hospital without ID physicians. Methods We conducted a retrospective observational analysis of data documented in clinical records by ward pharmacists in a 313-bed community hospital from April 2014 to March 2022. The ASP team conducted weekly reviews of targeted patients, provided feedback to physicians, and shared recommendations with ward pharmacists who then collaborated to optimize antimicrobial therapy. Using Python-based text mining with standardized technical terms and compound word extraction, we performed morphological analysis, co-occurrence network analysis, and hierarchical clustering to evaluate documentation patterns before and after ASP implementation in April 2018. Co-occurrence relationships were assessed using Dice coefficients (threshold, ≥ 0.3), and communities were detected using the Louvain algorithm. Changes in documentation patterns were compared using Fisher's exact test. Results The analysis included 1,353 pre-ASP and 5,155 post-ASP clinical records containing antimicrobial-related terms, which increased from 3.12–7.81% of the total pharmacy records. New strong co-occurrence relationships emerged in the post-ASP period for several laboratory parameters (c-reactive protein, 0.646; estimated glomerular filtration rate, 0.594; and white blood cell count, 0.582). Network analysis revealed a shift from medication-focused communities (Medication Review, Prescription Verification, and Patient Education) to infection-focused communities (Infection Assessment, Microbiological Review, and Severe Infection Management). Although Antimicrobial Management was consistently used in both periods (odds ratio [OR]: 0.70, 95% confidence interval [CI]: 0.38–1.20), cross-tabulation analysis increased significantly in Laboratory Monitoring (OR: 1.58, 95% CI: 1.39–1.78) and Infection Assessment (OR: 2.09, 95% CI: 1.85–2.36). Conclusions This pilot application of text mining demonstrated potential as a novel methodology for objectively evaluating qualitative changes in clinical practice patterns following ASP implementation, successfully identifying shifts in pharmacists' documentation focus and providing a foundation for future multi-center validation studies across diverse healthcare settings. antimicrobial stewardship program ward pharmacist text mining co-occurrence network analysis medical records infectious disease medium-sized hospital Figures Figure 1 Figure 2 Background Antimicrobial resistance (AMR) has emerged as a critical global health threat, leading to increased healthcare costs, treatment failure, and mortality rates [ 1 ]. The implementation of antimicrobial stewardship programs (ASPs) has been widely recognized as an essential strategy for optimizing antimicrobial use and combating AMR [ 2 ]. Guidelines such as those by the Infectious Diseases Society of America emphasize the importance of infectious disease (ID) physicians, along with specialized pharmacists, as core members of ASP teams [ 3 ], but the shortage of ID physicians poses significant challenges for healthcare institutions worldwide [ 4 ]. This shortage is particularly pronounced in Japan, where ID specialists are unevenly distributed among university hospitals and designated medical institutions [ 5 ], with medium-sized community hospitals in particular lacking sufficient resources to implement comprehensive ASPs. Ward pharmacists focusing on inpatient care make substantial contributions to the optimization of medication therapy across healthcare settings [ 6 ]. Their activities encompass a wide range of pharmaceutical care services, including medication order review, drug interaction assessment, medication safety monitoring, and providing medication-related consultation to healthcare professionals [ 7 ]. In the context of antimicrobial therapy, some studies have reported the successful involvement of ward pharmacists in specific aspects of antimicrobial stewardship, such as therapeutic drug monitoring-based dose optimization and renal function-based dose adjustment [ 8 , 9 ]. However, these interventions tend to be limited to specific clinical situations rather than being part of a systematic approach. Comprehensive antimicrobial stewardship activities are still typically conducted by dedicated antimicrobial stewardship pharmacists [ 10 , 11 ], and evidence regarding the systematic involvement of ward pharmacists in broad antimicrobial stewardship initiatives remains limited, partly due to the challenges in evaluating qualitative changes in clinical practice. Text mining analysis of medical records has emerged as a valuable tool for evaluating such changes in healthcare practices as well as professional behavior [ 12 ]. This approach enables quantitative analysis of qualitative data through techniques such as frequency analysis, co-occurrence network analysis, and cluster analysis [ 12 ]. It has been successfully applied in various medical fields, including the analysis of clinical decision-making patterns, assessment of healthcare quality improvements, and evaluation of interprofessional collaboration [ 13 , 14 ]. However, the application of text mining to evaluate antimicrobial stewardship activities and ward pharmacists' behavioral changes in antimicrobial management remains unexplored. Therefore, this study aimed to use text mining to analyze the changes in ward pharmacist antimicrobial management practices after the implementation of an ASP. Methods Aim and study design This was a single-center retrospective observational study. Its aim was to use Python-based text mining of clinical record text documented by ward pharmacists in a medium-sized Japanese hospital without ID physicians to analyze the qualitative changes in documentation patterns regarding antimicrobial stewardship practices after the implementation of an ASP. Setting In April 2018, we established a pharmacist-led ASP at our hospital, which had no ID physicians. The ASP team comprised nurses (0.1 full-time equivalent [FTE]), clinical microbiologists (0.2 FTE), pharmacists (1.0 FTE), and physicians without ID certification (0.1 FTE). The term "pharmacist-led" in our context refers to the fact that a dedicated ASP pharmacist, who had 3 years of clinical experience when the ASP was initiated in 2018 and consistently served in this role throughout the study period, functioned as the primary coordinator of the ASP activities. In addition to this dedicated ASP pharmacist, a dedicated drug information pharmacist with 15 years of clinical experience who was also a member of the infection control team participated in the weekly conferences. There were no changes in the ASP team membership or significant FTE variations throughout the study period. The intervention method employed was Prospective Audit and Feedback. The ASP team conducted weekly 2-hour conferences to review patients receiving antipseudomonal antibiotics, anti-methicillin-resistant Staphylococcus aureus drugs, prolonged antimicrobial therapy (≥ 14 days), or those with bloodstream infections. Prior to these conferences, the ASP pharmacist was responsible for identifying patients meeting intervention criteria and preparing comprehensive patient summaries. During these conferences, the team discussed the need for adjustments to treatment duration, additional laboratory tests, changes in antimicrobial agents, and optimization of dosing regimens. The ASP recommendations were communicated to attending physicians through electronic medical records and shared with ward pharmacists via email, with direct communication for urgent cases. The ward pharmacists then collaborated with attending physicians to optimize antimicrobial therapy based on these recommendations. Although an Infection Control Team had been active prior to April 2018, these ASP-related activities had not been previously conducted. Study data The study period (April 2014 to March 2022) was divided into two periods using April 2018 (when the ASP was implemented) as the cut-off point—the pre-ASP period (April 2014 to March 2018) and the post-ASP period (April 2018 to March 2022). Clinical records documented by ward pharmacists were extracted from the hospital's electronic medical record system (provided by Software Service, Inc.). Records containing the term "antimicrobial" (including alternative terms such as "antibiotic," "antibacterial," and "anti-infective") were included for analysis. Records documented by ASP pharmacists were categorized as "Team records" in our electronic medical record system and were excluded. Although no standardized templates were used among pharmacists during the study period, we cannot exclude the possibility that individual pharmacists may have used consistent phrasing in their documentation. We also collected data on the number of ward pharmacists and their years of clinical experience during both periods to assess potential confounding factors. Data collection and anonymization The data extracted from electronic medical records underwent a two-stage anonymization process. First, the medical information management department at Hirakata Kohsai Hospital replaced identifiable information (such as patient identifiers and medical record numbers) with anonymized codes. Second, the researchers transferred this data to a standalone environment and performed morphological analysis and part-of-speech filtering to detect and remove personal identifiers (such as organization names, personal names, and geographical locations) from the free-text portions. The dataset generated after this two-stage anonymization process was stored in a password-protected Microsoft OneDrive. The original data and anonymization correspondence table were securely managed within the hospital. Text preprocessing All analyses were performed using Python 3.9.4. We performed morphological analysis using mecab-python3 (version 1.0.10) with the ipadic dictionary and part-of-speech filtering using the unicodedata2 package (version 16.0.0) on the anonymized dataset. The morphological analysis was executed in the -Ochasen mode for detailed part-of-speech tagging. Parts of speech were classified into nouns, verbs, adjectives, adverbs, interjections, and prenominal adjectives. To extract meaningful words, we excluded certain word categories—suffixes, numbers, pronouns, auxiliary words, and formal nouns from nouns; auxiliary adjectives from adjectives; and all symbols, particles, and auxiliary verbs. Standardization of technical terms Compound terms were extracted using the termextract package (version 0.12b) in Python. Terms were selected based on their importance scores calculated using the log-frequency ratio method (lr_mode = 1, average_rate = True). Words with scores in the top 5% were classified as compound words. To standardize terminology variations, we used the Medical Subject Headings (MeSH) and Systematized Nomenclature of Medicine (SNOMED) controlled vocabularies. Terms were standardized according to the following procedure: Two pharmacists with dedicated ASP experience collaboratively searched for equivalent MeSH terms for each extracted term using the MeSH Browser ( https://meshb.nlm.nih.gov/ ). When an appropriate MeSH term was found, the extracted term was replaced with the standardized MeSH term. If no suitable MeSH term was available, the pharmacists searched for an equivalent SNOMED term using the SNOMED CT Browser ( https://browser.ihtsdotools.org/ ). Terms that were mapped to the same MeSH or SNOMED identifier (e.g., D000900 and 404684003, respectively) were treated as synonyms to ensure consistent terminology. In cases where the pharmacists encountered difficulty in term mapping, consensus was reached through discussion. Single alphabet characters and units were excluded as meaningful analysis targets. The dataset with these defined compound words, synonyms, and exclusion words was then reprocessed using morphological analysis to obtain the final processed dataset. Co-occurrence analysis We constructed a co-occurrence network centered on the term "antimicrobial" using the networkx package (version 3.2.1) in Python. As "antimicrobial" appeared in all records and was expected to show large differences in detection frequency with other terms, we used the Dice coefficient to evaluate the strength of co-occurrence relationships. Considering that the Dice coefficient typically yields higher values than the Jaccard coefficient, we set a conservative threshold of ≥ 0.3 to define the presence of co-occurrence relationships. The network was exported in GraphML format and visualized using gephi (version 0.10.1). To ensure network readability, we applied the ForceAtlas layout algorithm, adjusted edge thickness based on the Dice coefficient values, and adjusted node sizes according to their frequency of occurrence. We have depicted only the minimum spanning tree, prioritizing edges with higher coefficients. Community detection within the co-occurrence network was performed using the Louvain algorithm implemented in the python-louvain package (version 0.16). The best partition was determined using modularity optimization (community.best_partition(G)), which identifies groups of densely connected terms. For the interpretation of the communities, two pharmacists with dedicated ASP experience independently reviewed the terms within each community and assigned names reflecting the primary focus of the grouped terms. Any disagreements were resolved through discussion with a third researcher (a faculty member with experience in hospital infection control). The detailed naming process and term lists for each community are provided in the supplementary materials. Cluster analysis Hierarchical clustering was performed using Ward's method in the scipy package (version 1.13.1) on terms detected as nodes in the co-occurrence network. Dissimilarity was initially defined as 1- Dice coefficient, but the final cluster distances were computed using Ward's method, which minimizes the variance increase at each merging step. Based on preliminary examinations of the dendrogram structure, the number of clusters was set to 10% of the total detected terms to balance interpretability. The naming process for the clusters followed the same procedure as described in the co-occurrence analysis section. To evaluate changes in documentation content before and after ASP implementation, we performed cross-tabulation comparing the frequency of occurrence between the pre-ASP and post-ASP periods for the obtained clusters. To exclude the potential impact of the coronavirus disease 2019 (COVID-19) pandemic on our findings, we conducted a sensitivity analysis that excluded data from April 2020 onward (the beginning of the pandemic's impact in Japan). Odds ratios and their 95% confidence intervals were calculated; statistical comparisons between the periods were conducted using Fisher's exact test. Statistical significance was set at p < 0.05. Data management To ensure reproducibility, the data extraction conditions, and analysis programs were documented in the electronic medical record system. As specified in the data management policy established by the research ethics committee, the original non-anonymized data were designated to be retained at Hirakata Kohsai Hospital for 3 years from the study completion date, and the processed dataset at Kyoto Pharmaceutical University for 5 years. The policy also specified that all data must be destroyed using methods that prevent restoration after these retention periods. Results Basic characteristics of the text data The basic characteristics of the study data are presented in Table 1. Ward pharmacist staffing remained relatively stable throughout the study period, with a slight increase in the annual average number after ASP implementation (12.8 to 14.5) and consistent average years of clinical experience (7.3 to 7.4 years). The proportion of records containing the term "antimicrobial" increased substantially, from 1,353 (3.12%) in the pre-ASP period to 5,155 (7.81%) in the post-ASP period. The preprocessing identified 29 compound words, 20 sets of synonyms, and 37 exclusion words (Additional file 1, Supplementary Tables 1 and 2). The morphological analysis resulted in the extraction of 75,148 words (55.5 words per record) in the pre-ASP period and 338,134 words (65.6 words per record) in the post-ASP period. Table 1 Basic characteristics of the study data Pre-ASP Post-ASP Ward pharmacist staffing Number of ward pharmacists at period start 11 13 Annual average number of ward pharmacists 12.8 ± 2.4 14.5 ± 1.3 New pharmacists during the period 9 8 Departing pharmacists during the period 7 6 Average years of clinical experience 7.3 ± 7.8 7.4 ± 5.7 Documentation characteristics Total clinical records 43,374 65,982 Records containing "antimicrobial" 1,353 (3.12%) 5,155 (7.81%) Average characters per record 267.1 ± 116.9 353.1 ± 209.3 Average raw terms per record a 149.1 ± 67.2 197.1 ± 120.5 Average analyzed terms per record b 55.5 ± 24.0 65.6 ± 37.7 Values are presented as mean ± standard deviation where applicable. a Raw terms indicate terms extracted after morphological analysis before filtering. b Analyzed terms indicate terms used for final analysis after morphological analysis, part-of-speech filtering, and standardization. Abbreviations: ASP, antimicrobial stewardship program. Terms co-occurring with "antimicrobial" Terms with direct co-occurrence relationships (Dice coefficient ≥0.3) with "antimicrobial" were identified—27 terms in the pre-ASP period and 31 terms in the post-ASP period (Table 2). Three terms showed strong co-occurrence relationships (Dice coefficient ≥0.5) in both periods—"Laboratory Test" (Dice coefficient: pre-ASP, 0.525; post-ASP, 0.672), "Creatinine" (pre-ASP, 0.507; post-ASP, 0.609), and "Verification" (pre-ASP, 0.681; post-ASP, 0.584). In the pre-ASP period, "Necessity" and "Particularly" were characteristically detected, while in the post-ASP period, new strong relationships emerged with "C-Reactive Protein (CRP)", "estimated Glomerular Filtration Rate (eGFR)", "White Blood Cell Count (WBC)", and "Aspartate Aminotransferase (AST)". Table 2 Terms with strong co-occurrence relationships with "antimicrobials" (Dice coefficient ≥ 0.3) Pre-ASP (n = 1,353) Post-ASP (n = 5,155) Term Frequency a Co-occurrence b Frequency a Co-occurrence b ALT 1,677 (32.5) 0.491 AST 1,740 (33.8) 0.505 Administration 282 (20.8) 0.345 1,378 (26.7) 0.422 Bacterial Test 936 (18.2) 0.307 CRP 338 (25.0) 0.400 2,457 (47.7) 0.646 Caution 420 (31.0) 0.474 1,114 (21.6) 0.355 Ceftriaxone 975 (18.9) 0.318 Continuation 1,041 (20.2) 0.336 Creatinine 459 (33.9) 0.507 2,258 (43.8) 0.609 Decrease 1,056 (20.5) 0.340 Discontinuation 946 (18.4) 0.310 Dosage 296 (21.9) 0.359 Drugs 428 (31.6) 0.481 1,098 (21.3) 0.351 eGFR 366 (27.1) 0.426 2,179 (42.3) 0.594 Explanation 378 (27.9) 0.437 Fever 1,044 (20.3) 0.337 Guidance 364 (26.9) 0.424 Hospitalization 1,014 (19.7) 0.329 Initiation 358 (26.5) 0.419 1,270 (24.6) 0.395 Laboratory Test 482 (35.6) 0.525 2,609 (50.6) 0.672 Management Method 281 (20.8) 0.344 Med. Adherence 353 (26.1) 0.414 Med. Reconciliation 310 (22.9) 0.373 916 (17.8) 0.302 Med. Taking 276 (20.4) 0.339 931 (18.1) 0.306 Medicine 370 (27.3) 0.430 Modification 315 (23.3) 0.378 1,424 (27.6) 0.433 Necessity 618 (45.7) 0.627 1,025 (19.9) 0.332 Neutrophil 1,543 (29.9) 0.461 Oral Administration 327 (24.2) 0.389 1,344 (26.1) 0.414 Particularly 497 (36.7) 0.537 Prescription 384 (28.4) 0.442 1,258 (24.4) 0.392 Qualitative Test 1,497 (29.0) 0.450 Renal Function 346 (25.6) 0.407 Report 266 (19.7) 0.329 SBT/ABPC 1,062 (20.6) 0.342 Safety 434 (32.1) 0.486 Tablet 387 (28.6) 0.445 1,260 (24.4) 0.393 Take 333 (24.6) 0.395 979 (19.0) 0.319 Tend 1,158 (22.5) 0.367 Verification 698 (51.6) 0.681 2,124 (41.2) 0.584 WBC 2,113 (41.0) 0.582 a Numbers in parentheses indicate the percentage of records containing the corresponding term b Dice coefficients indicate the strength of co-occurrence relationships Blank cells indicate that the term was not detected or did not meet the threshold for co-occurrence (Dice coefficient ≥ 0.3) in that period. Abbreviations: ALT, alanine aminotransferase; ASP, antimicrobial stewardship program; AST, aspartate aminotransferase; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; Med., medication; SBT/ABPC, sulbactam/ampicillin; WBC, white blood cell count Network analysis of co-occurring terms A total of 84 co-occurring terms formed a network, with 66 nodes in the pre-ASP period and 63 nodes in the post-ASP period (Figure 1). Louvain algorithm analysis detected seven communities in each period (Additional file 1, Supplementary Table 3). Four communities were common to both periods: Laboratory Monitoring, Antimicrobial Management, Information Collection, and Dose Optimization. The pre-ASP period was characterized by communities related to Medication Review, Prescription Verification, and Patient Education, while the post-ASP period showed newly emerged communities related to Infection Assessment, Microbiological Review, and Severe Infection Management. Cluster analysis and cross-tabulation Ward’s method of hierarchical cluster analysis of the 84 terms identified in the network analysis yielded seven groups (Figure 2). Cross-tabulation analysis comparing the pre- and post-ASP periods showed significant increases in Laboratory Monitoring and Infection Assessment and significant decreases for Prescription Verification, Medication Review, Patient Education, and Information Collection. Only Antimicrobial Management remained consistently high in both periods (Table 3). Sensitivity analysis comparing the pre-ASP period with only the early post-ASP period (April 2018 to March 2020, n=1,400) showed similar patterns to the main analysis, with significant increases in Laboratory Monitoring (OR: 1.79, 95% CI: 1.54–2.09, p<0.01) and Infection Assessment (OR: 2.05, 95% CI: 1.75–2.39, p<0.01) (Additional file 1, Supplementary Table 4). Table 3 Changes in documentation patterns before and after ASP implementation Pre-ASP (n = 1,353) Post-ASP (n = 5,155) Odds ratio (95% CI) p value Antimicrobial Management 1,337 (98.8) 5,068 (98.3) 0.70 (0.38 − 1.20) 0.22 Laboratory Monitoring 617 (45.6) 2,934 (56.9) 1.58 (1.39 − 1.78) < 0.01 Infection Assessment 598 (44.2) 3,214 (62.4) 2.09 (1.85 − 2.36) < 0.01 Information Collection 494 (36.5) 1,405 (27.3) 0.65 (0.57 − 0.74) < 0.01 Prescription Verification 746 (55.1) 1,896 (36.8) 0.47 (0.42 − 0.54) < 0.01 Medication Review 992 (73.3) 3,264 (63.3) 0.63 (0.55 − 0.72) < 0.01 Patient Education 591 (43.7) 917 (17.8) 0.28 (0.24 − 0.32) < 0.01 Values are expressed as n (%). Statistical comparisons were performed using Fisher's exact test. p values < 0.05 were considered statistically significant. Abbreviations: ASP, antimicrobial stewardship program; CI, confidence interval Discussion This study used text mining analysis to evaluate changes in ward pharmacist interventions for antimicrobial therapy before and after the implementation of an ASP. The analysis revealed characteristic changes in documentation patterns—specifically, an increase in ID-related terms and a relative decrease in medication guidance-related terms. Network analysis demonstrated a shift from communities related to Medication Review, Prescription Verification, and Patient Education in the pre-ASP period to those focused on Microbiological Review, Infection Assessment, and Severe Infection Management in the post-ASP period. In the context of antimicrobial therapy, CRP level and WBC are essential indicators for assessing infection severity and treatment efficacy [15]; moreover, renal function and liver function test results (eGFR and AST level, respectively) are crucial for dose adjustment [16, 17], and microbiological test results are essential for appropriate antimicrobial selection and improving treatment outcomes [18, 19]. In our study, new co-occurrence relationships involving these specific laboratory parameters emerged in the post-ASP period, beyond the general association with Laboratory Tests in the pre-ASP period. Additionally, the emergence of the Microbiological Review community suggests that ward pharmacists may have evolved their practices to include more specialized therapeutic support, particularly in terms of pathogen-based antimicrobial selection. The basic roles of ward pharmacists include medication management activities such as safety verification and providing medication guidance [7, 20, 21]. In our analysis, while terms related to Microbiological Review and Infection Assessment increased after ASP implementation, the proportion of terms related to Medication Review, Patient Education, and Prescription Verification decreased. However, the overall clinical documentation by pharmacists also increased 1.5-fold after ASP implementation, with antimicrobial-related records showing a particularly marked increase (from 3.12% to 7.81% of the total records). These findings suggest that ward pharmacists expanded their expertise into antimicrobial stewardship while maintaining their basic medication management activities. Previous studies evaluating ASP effectiveness focused primarily on large hospitals with adequate staff, including one to two ID physicians, at least one dedicated pharmacist, and one or more infection control nurses [10, 22, 23]. In contrast, even among legally designated medical institutions for IDs in Japan, only 33.3% have ID specialists [5], and many hospitals struggle to allocate dedicated ASP staff owing to resource constraints [24]. Although several reports have described the feasibility and effectiveness of ASPs in resource-limited settings [25–27], very few studies have examined how ward pharmacists adapt their practices to support antimicrobial stewardship in hospitals without ID specialists. Our text mining analysis provides objective evidence on how ward pharmacists can evolve their practice patterns to support antimicrobial stewardship in resource-limited settings. Conventional evaluations of ASP activities have primarily focused on quantitative indicators such as antimicrobial use and treatment duration [28]. Although these measures are essential for assessing program effectiveness, they may not fully capture the evolution of ward pharmacist practices and their collaboration with ASP teams. Text mining of medical records has emerged as a complementary analytical approach that can reveal changes in clinical practice patterns [29, 30]. In this study, this analytical method provided unique insights into how ward pharmacists adapt their roles and responsibilities in response to new antimicrobial stewardship initiatives. Nonetheless, this study had some limitations. First, text mining cannot be used to establish causal relationships, making it impossible to determine whether the changes in documentation patterns resulted from improved pharmacist expertise or ASP interventions. Although our data showed that the average years of clinical experience of ward pharmacists remained relatively stable throughout the study period (pre-ASP 7.3 years vs. post-ASP 7.4 years), we cannot exclude the possibility that changes in experience levels influenced the results. Additionally, while our sensitivity analysis showed similar results when excluding the COVID-19 pandemic period, we cannot rule out the possibility that major events that significantly impact awareness of IDs, such as the pandemic, may have influenced our findings. Second, as this was a single-center retrospective study, the generalizability of our findings remains limited. However, the primary aim of this study was to qualitatively evaluate antimicrobial stewardship activities using a novel text mining approach. We recognize that this investigation serves as a pilot study proposing methodologies for handling unprocessed electronic medical record data containing personal information in text mining research and for evaluating data after morphological analysis. Future studies applying this methodology across healthcare institutions with diverse backgrounds and examining its applicability to other healthcare professionals would contribute to building more robust evidence. Conclusions As a pilot study, we applied a novel text mining approach to qualitatively evaluate antimicrobial stewardship activities in a single Japanese hospital. Future research applying this methodology across diverse healthcare institutions and examining its applicability to other healthcare professionals will contribute to building more robust evidence for the evolution of pharmacy practice in antimicrobial stewardship. Abbreviations ADR: adverse drug reaction, ALT: alanine aminotransferase, AMR: antimicrobial resistance, AST: aspartate aminotransferase, ASP: antimicrobial stewardship program, CCr: creatinine clearance, CI: confidence interval, COVID-19: coronavirus disease-2019, CRP: C-reactive protein, eGFR: estimated glomerular filtration rate, FTE: full-time equivalent, ID: infectious disease, Med.: medication, MeSH: Medical Subject Headings, MRSA: methicillin-resistant Staphylococcus aureus , OR: odds ratio, OTC: over-the-counter, Pharm.: pharmaceutical, SBT/ABPC: sulbactam/ampicillin, SNOMED: Systematized Nomenclature of Medicine, WBC: white blood cell count Declarations Ethics approval and consent to participate This study was approved by the Ethics Committees of Hirakata Kohsai Hospital (approval number: 2022-009) and Kyoto Pharmaceutical University (approval number: E23-029). The study was conducted in accordance with the Declaration of Helsinki and the Ethical Guidelines for Life Science and Medical Research Involving Human Subjects. An opt-out approach was used, where information about the study was made available on the hospital website, and participants were provided the opportunity to refuse participation. Consent for publication Not applicable. Availability of data and materials The datasets generated and/or analyzed during the current study are not publicly available due to privacy concerns and institutional policies regarding medical records but are available from the corresponding author on reasonable request with approval from the Ethics Committees of Hirakata Kohsai Hospital and Kyoto Pharmaceutical University. The detailed protocol for text preprocessing, including the standardized terminology list and compound word definitions, is available in the supplementary materials. Competing interests RI received grants from Kowa Company Ltd., the Japan Pharmaceutical Association, and the Japan Society for the Promotion of Science. YM received grants from Pfizer Japan Inc., Kowa Company Ltd., the Japan Pharmaceutical Association, the Japan Society for the Promotion of Science, and the Ministry of Health, Labour and Welfare. YM is also a board member of the Japanese Society of Pharmaceutical Health Care and Sciences and the Japanese Society for Infection Prevention and Control and a committee member of the AMR Clinical Reference Center. The other authors have no conflicts of interest to declare. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Authors' contributions KS designed the study protocol, led the antimicrobial stewardship program, collected and analyzed the data, performed the MeSH and SNOMED term mapping, participated in community naming, and drafted the manuscript. SK and RI critically reviewed the manuscript. YM supervised the research project, participated in community naming, and critically reviewed the manuscript. All authors read and approved the final version of the manuscript. Acknowledgements We thank the staff of the Infection Control Committee and the Pharmacy Department at the Federation of National Public Service Personnel Mutual Aid Associations Hirakata Kohsai Hospital for managing the ASP. We also thank Hikaru Nishiyama for his contributions to ASP implementation, MeSH and SNOMED term mapping, and community naming. We are grateful to the Medical Information Management Department at the Federation of National Public Service Personnel Mutual Aid Associations Hirakata Kohsai Hospital for their assistance with data extraction and anonymization. References Cosgrove SE. The relationship between antimicrobial resistance and patient outcomes: mortality, length of hospital stay, and health care costs. Clin Infect Dis . 2006;42 Suppl 2:S82-9. doi: 10.1086/499406. Doron S, Davidson LE. Antimicrobial stewardship. Mayo Clin Proc . 2011;86(11):1113-23. doi: 10.4065/mcp.2011.0358. Barlam TF, Cosgrove SE, Abbo LM, MacDougall C, Schuetz AN, Septimus EJ, et al. Implementing an antibiotic stewardship program: guidelines by the infectious diseases society of America and the society for healthcare epidemiology of America. Clin Infect Dis . 2016;62(10):e51-77. doi: 10.1093/cid/ciw118. Hussaini SMQ. Dearth of infectious diseases physicians as the USA faces a global pandemic. Lancet Infect Dis . 2020;20(6):648-9. doi: 10.1016/S1473-3099(20)30377-7. Hagiya H. Shortage and unequal distribution of infectious disease specialists in Japan: how can we refine the current situation? PLOS One . 2023;18(10):e0291677. doi: 10.1371/journal.pone.0291677. Chen K-L, Hunag C-F, Sheng W-H, Chen Y-K, Wang C-C, Shen L-J. Impact of integrated medication management program on medication errors in a medical center: an interrupted time series study. BMC Health Serv Res . 2022;22(1):796. doi: 10.1186/s12913-022-08178-w. Iqbal MJ, Mohammad Ishaq G, Assiri AA. Connecting pharmacists and other health care providers (HCPs) towards drug therapy optimization: A pharmaceutical care approach. Int J Clin Pract . 2023;2023:3336736. doi: 10.1155/2023/3336736. Oda K, Saito H, Jono H. Bayesian prediction-based individualized dosing of anti-methicillin-resistant Staphylococcus aureus treatment: recent advancements and prospects in therapeutic drug monitoring. Pharmacol Ther . 2023;246:108433. doi: 10.1016/j.pharmthera.2023.108433. Sukkha S, Rattanavipanon W, Chamroenwit B, Sanganurak M, Nathisuwan S, Chaikledkaew U, Kongwatcharapong J. Quality assessment and cost saving of renal dosing recommendation by clinical pharmacists at medical wards in Thailand. Int J Clin Pharm . 2020;42(2):610-6. doi: 10.1007/s11096-020-01016-1. Okada N, Azuma M, Tsujinaka K, Abe A, Takahashi M, Yano Y, et al. Clinical impact of a pharmacist-driven prospective audit with intervention and feedback on the treatment of patients with bloodstream infection. Antibiotics (Basel) . 2022;11(9):1144. doi: 10.3390/antibiotics11091144. Nampoothiri V, Hisham M, Mbamalu O, Mohamed ZU, Singh SK, Charani E. Evolution of pharmacist roles in antimicrobial stewardship: A 20-year systematic review. Int J Infect Dis . 2025;151:107306. doi: 10.1016/j.ijid.2024.107306. Percha B. Modern clinical text mining: A guide and review. Annu Rev Biomed Data Sci . 2021;4:165-87. doi: 10.1146/annurev-biodatasci-030421-030931. van de Burgt BWM, Wasylewicz ATM, Dullemond B, Grouls RJE, Egberts TCG, Bouwman A, Korsten EMM. Combining text mining with clinical decision support in clinical practice: A scoping review. J Am Med Inform Assoc . 2023;30(3):588-603. doi: 10.1093/jamia/ocac240. Grob M, Jenders RA, Rappelsberger A, Adlassnig K-P. Integrated perspectives on clinical decision support: A comparative analysis of knowledge management approaches. Stud Health Technol Inform . 2024;316:1822-6. doi: 10.3233/SHTI240785. Muna AM, Alhameed RA. The role of C-reactive protein and white blood cell count as diagnostic, prognostic, and monitoring markers in bacterial orofacial infections. J Oral Maxillofac Surg . 2022;80(3):530-6. doi: 10.1016/j.joms.2021.08.158. Keller F, Schröppel B, Ludwig U. Pharmacokinetic and pharmacodynamic considerations of antimicrobial drug therapy in cancer patients with kidney dysfunction. World J Nephrol . 2015;4(3):330-44. doi: 10.5527/wjn.v4.i3.330. Halilovic J, Heintz BH. Antibiotic dosing in cirrhosis. Am J Health Syst Pharm . 2014;71(19):1621-34. doi: 10.2146/ajhp140031. Hamilton WL, Pires S-M, Lippett S, Gudka V, Cross ELA, Llewelyn MJ. The impact of diagnostic microbiology on de-escalation of antimicrobial therapy in hospitalised adults. BMC Infect Dis . 2020;20(1):102. doi: 10.1186/s12879-020-4823-4. Stelling J, Read JS, Fritch W, O’Brien TF, Peters R, Clark A, et al. Surveillance of antimicrobial resistance and evolving microbial populations in Vermont: 2011-2018. Expert Rev Anti Infect Ther . 2020;18(10):1055-62. doi: 10.1080/14787210.2020.1776114. Salmasi S, Tsao NW, Li K, Shaske JN, Marra CA, Lynd LD. Characterization of pharmacist-based medication management services in a community pharmacy. Res Social Adm Pharm . 2020;16(2):178-82. doi: 10.1016/j.sapharm.2019.04.051. Bainbridge J, Barnhart R, Fuller R, Hellerslia VT, Kidd J, Merrill S, et al. The role of clinical pharmacists in patient-centric comprehensive multiple sclerosis care. Int J MS Care . 2024;26(1):1-7. doi: 10.7224/1537-2073.2022-051. Nathwani D, Varghese D, Stephens J, Ansari W, Martin S, Charbonneau C. Value of hospital antimicrobial stewardship programs [ASPs]: A systematic review. Antimicrob Resist Infect Control . 2019;8:35. doi: 10.1186/s13756-019-0471-0. Giamarellou H, Galani L, Karavasilis T, Ioannidis K, Karaiskos I. Antimicrobial stewardship in the hospital setting: A narrative review. Antibiotics (Basel) . 2023;12(10):1557. doi: 10.3390/antibiotics12101557. Maeda M, Muraki Y, Kosaka T, Yamada T, Aoki Y, Kaku M, et al. Essential human resources for antimicrobial stewardship teams in Japan: estimates from a nationwide survey conducted by the Japanese Society of Chemotherapy. J Infect Chemother . 2019;25(9):653-6. doi: 10.1016/j.jiac.2019.05.012. Banerjee S, Gupta N, Ray Y, Kodan P, Khot WY, Fazal F, et al. Impact of trainee-driven Antimicrobial Stewardship Program in a high burden resource-limited setting. Infez Med . 2020;28(3):367-72. https:// pubmed.ncbi.nlm.nih.gov/ 32920572. Sawada K, Inose R, Goto R, Nakatani T, Kono S, Muraki Y. Effectiveness of noncertified pharmacist-led antimicrobial stewardship programs in a medium-sized hospital without an infectious disease specialist: A retrospective pre-post study. Open Forum Infect Dis . 2023;10(3):ofad116. doi: 10.1093/ofid/ofad116. Nakano Y, Hirai T, Murata M, Yasukochi H, Ura K, Sueyasu Y, et al. Impact of pharmacist-driven antimicrobial stewardship interventions in a secondary care facility in Japan: A pragmatic quasi-experimental study. J Infect Chemother . 2025;31(1):102503. doi: 10.1016/j.jiac.2024.08.018. Septimus E. Antimicrobial stewardship-qualitative and quantitative outcomes: the role of measurement. Curr Infect Dis Rep . 2014;16(11):433. doi: 10.1007/s11908-014-0433-x. Marafino BJ, Park M, Davies JM, Thombley R, Luft HS, Sing DC, et al. Validation of prediction models for critical care outcomes using natural language processing of electronic health record data. JAMA Netw Open . 2018;1(8):e185097. doi: 10.1001/jamanetworkopen.2018.5097. Wieland-Jorna Y, van Kooten D, Verheij RA, de Man Y, Francke AL, Oosterveld-Vlug MG. Natural language processing systems for extracting information from electronic health records about activities of daily living. A systematic review. JAMIA Open . 2024;7(2):ooae044. doi: 10.1093/jamiaopen/ooae044. Additional Declarations Competing interest reported. RI received grants from Kowa Company Ltd., the Japan Pharmaceutical Association, and the Japan Society for the Promotion of Science. YM received grants from Pfizer Japan Inc., Kowa Company Ltd., the Japan Pharmaceutical Association, the Japan Society for the Promotion of Science, and the Ministry of Health, Labour and Welfare. YM is also a board member of the Japanese Society of Pharmaceutical Health Care and Sciences and the Japanese Society for Infection Prevention and Control and a committee member of the AMR Clinical Reference Center. The other authors have no conflicts of interest to declare. Supplementary Files TextMiningJPHCSR1Sup20250331clean.docx Additional file 1 Supplementary Table 1. Term standardization and synonyms used in the text mining analysis Supplementary Table 2. Terms excluded from the text mining analysis Supplementary Table 3. Communities and clusters identified in the text mining analysis Supplementary Table 4. Changes in documentation patterns before and after ASP implementation: Sensitivity analysis excluding the COVID-19 pandemic period Cite Share Download PDF Status: Published Journal Publication published 23 Apr, 2025 Read the published version in Journal of Pharmaceutical Health Care and Sciences → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6199166","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":437684282,"identity":"1e3ccdda-ff29-4cd5-9970-b37d3316ecd3","order_by":0,"name":"Keisuke Sawada","email":"","orcid":"","institution":"Federation of National Public Service Personnel Mutual Aid Associations Hirakata Kohsai Hospital","correspondingAuthor":false,"prefix":"","firstName":"Keisuke","middleName":"","lastName":"Sawada","suffix":""},{"id":437684283,"identity":"480781d6-3ac1-41ff-af32-79d1537e3ec1","order_by":1,"name":"Shuji Kono","email":"","orcid":"","institution":"Federation of National Public Service Personnel Mutual Aid Associations Hirakata Kohsai Hospital","correspondingAuthor":false,"prefix":"","firstName":"Shuji","middleName":"","lastName":"Kono","suffix":""},{"id":437684284,"identity":"96300f8f-1b2e-42fc-a884-5f661a127255","order_by":2,"name":"Ryo Inose","email":"","orcid":"","institution":"Kyoto Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Ryo","middleName":"","lastName":"Inose","suffix":""},{"id":437684285,"identity":"b7a23131-240e-4998-af3b-d70e345b72ec","order_by":3,"name":"Yuichi Muraki","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABFElEQVRIiWNgGAWjYPCCBB5+Bh4GxgYIjw0qegC/FskGiBYJorUwGBzA1IIdGBzgTpOuqEmTMb529gDjzDabOn4G3mOPeRjs5BkYz2K1xuAA7zbJM8dyeMxu5yUwbmxLk5Bs4Es35mFINmxgOJeAVcv9t9skG9gqgFpyDBgfth2WMLj/xkyah4EZqPyMAU5bGv5V8BjPhmk5wAPSUo9fS2NbDo+BNFDLRoSWwzi1SB7g3WzZ2JfGIwH0y8EZ59IkZzbwpUnOMThu2IbDL3wHeDfebPiWbM8/O/fgw54yG35QiEm8qaiW55fAHmIKyKIHGMExwgNyMDB2JM5g08Eg34DC/QPTAgL8PVi1jIJRMApGwYgDAF1VW1BpcsPmAAAAAElFTkSuQmCC","orcid":"","institution":"Kyoto Pharmaceutical University","correspondingAuthor":true,"prefix":"","firstName":"Yuichi","middleName":"","lastName":"Muraki","suffix":""}],"badges":[],"createdAt":"2025-03-11 01:23:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6199166/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6199166/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40780-025-00439-0","type":"published","date":"2025-04-23T15:57:31+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":79909826,"identity":"247a3ca9-9bca-4377-979e-e83488210e3a","added_by":"auto","created_at":"2025-04-04 11:25:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1647527,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCo-occurrence network analysis of terms related to antimicrobials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAbbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; ASP, antimicrobial stewardship program; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; Med., medication; SBT/ABPC, sulbactam/ampicillin; WBC, white blood cell count\u003c/p\u003e","description":"","filename":"TextMiningJPHCSFig120250331.png","url":"https://assets-eu.researchsquare.com/files/rs-6199166/v1/20df76084c5ed8dbab25faae.png"},{"id":79908205,"identity":"b9f263a9-fdc9-4e32-b289-972830b71e7e","added_by":"auto","created_at":"2025-04-04 11:17:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":499637,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHierarchical cluster analysis of terms related to antimicrobials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe value 2.9 represents the maximum cluster distance between groups.\u003c/p\u003e\n\u003cp\u003eAbbreviations: ADR, adverse drug reaction; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CCr, creatinine clearance; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; Med., medication; OTC, over-the-counter; Pharm., pharmaceutical; SBT/ABPC, sulbactam/ampicillin; WBC, white blood cell count\u003c/p\u003e","description":"","filename":"TextMiningJPHCSFig220250331.png","url":"https://assets-eu.researchsquare.com/files/rs-6199166/v1/bc9f4eecdf8b1754aa3e2303.png"},{"id":81569631,"identity":"5cf9ee50-f312-41bf-b140-f0f0f0fa8eb8","added_by":"auto","created_at":"2025-04-28 16:08:21","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2670486,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6199166/v1/6c7d6d6b-d0d8-4bb1-b255-50f78e61590b.pdf"},{"id":79908209,"identity":"25553de2-77e9-47bc-a4dd-fd4d71252c2e","added_by":"auto","created_at":"2025-04-04 11:17:45","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":43438,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 1\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSupplementary Table 1. Term standardization and synonyms used in the text mining analysis\u003c/p\u003e\n\u003cp\u003eSupplementary Table 2. Terms excluded from the text mining analysis\u003c/p\u003e\n\u003cp\u003eSupplementary Table 3. Communities and clusters identified in the text mining analysis\u003c/p\u003e\n\u003cp\u003eSupplementary Table 4. Changes in documentation patterns before and after ASP implementation: Sensitivity analysis excluding the COVID-19 pandemic period\u003c/p\u003e","description":"","filename":"TextMiningJPHCSR1Sup20250331clean.docx","url":"https://assets-eu.researchsquare.com/files/rs-6199166/v1/6b9714d326d8e89870b31765.docx"}],"financialInterests":"Competing interest reported. RI received grants from Kowa Company Ltd., the Japan Pharmaceutical Association, and the Japan Society for the Promotion of Science. \nYM received grants from Pfizer Japan Inc., Kowa Company Ltd., the Japan Pharmaceutical Association, the Japan Society for the Promotion of Science, and the Ministry of Health, Labour and Welfare. YM is also a board member of the Japanese Society of Pharmaceutical Health Care and Sciences and the Japanese Society for Infection Prevention and Control and a committee member of the AMR Clinical Reference Center. \nThe other authors have no conflicts of interest to declare.","formattedTitle":"Qualitative Changes in Clinical Records After Implementation of Pharmacist-Led Antimicrobial Stewardship Program: A Text Mining Analysis","fulltext":[{"header":"Background","content":"\u003cp\u003eAntimicrobial resistance (AMR) has emerged as a critical global health threat, leading to increased healthcare costs, treatment failure, and mortality rates [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. The implementation of antimicrobial stewardship programs (ASPs) has been widely recognized as an essential strategy for optimizing antimicrobial use and combating AMR [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Guidelines such as those by the Infectious Diseases Society of America emphasize the importance of infectious disease (ID) physicians, along with specialized pharmacists, as core members of ASP teams [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], but the shortage of ID physicians poses significant challenges for healthcare institutions worldwide [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. This shortage is particularly pronounced in Japan, where ID specialists are unevenly distributed among university hospitals and designated medical institutions [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], with medium-sized community hospitals in particular lacking sufficient resources to implement comprehensive ASPs.\u003c/p\u003e \u003cp\u003eWard pharmacists focusing on inpatient care make substantial contributions to the optimization of medication therapy across healthcare settings [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Their activities encompass a wide range of pharmaceutical care services, including medication order review, drug interaction assessment, medication safety monitoring, and providing medication-related consultation to healthcare professionals [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. In the context of antimicrobial therapy, some studies have reported the successful involvement of ward pharmacists in specific aspects of antimicrobial stewardship, such as therapeutic drug monitoring-based dose optimization and renal function-based dose adjustment [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. However, these interventions tend to be limited to specific clinical situations rather than being part of a systematic approach. Comprehensive antimicrobial stewardship activities are still typically conducted by dedicated antimicrobial stewardship pharmacists [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], and evidence regarding the systematic involvement of ward pharmacists in broad antimicrobial stewardship initiatives remains limited, partly due to the challenges in evaluating qualitative changes in clinical practice.\u003c/p\u003e \u003cp\u003eText mining analysis of medical records has emerged as a valuable tool for evaluating such changes in healthcare practices as well as professional behavior [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. This approach enables quantitative analysis of qualitative data through techniques such as frequency analysis, co-occurrence network analysis, and cluster analysis [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. It has been successfully applied in various medical fields, including the analysis of clinical decision-making patterns, assessment of healthcare quality improvements, and evaluation of interprofessional collaboration [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. However, the application of text mining to evaluate antimicrobial stewardship activities and ward pharmacists' behavioral changes in antimicrobial management remains unexplored. Therefore, this study aimed to use text mining to analyze the changes in ward pharmacist antimicrobial management practices after the implementation of an ASP.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eAim and study design\u003c/h2\u003e \u003cp\u003eThis was a single-center retrospective observational study. Its aim was to use Python-based text mining of clinical record text documented by ward pharmacists in a medium-sized Japanese hospital without ID physicians to analyze the qualitative changes in documentation patterns regarding antimicrobial stewardship practices after the implementation of an ASP.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSetting\u003c/h3\u003e\n\u003cp\u003eIn April 2018, we established a pharmacist-led ASP at our hospital, which had no ID physicians. The ASP team comprised nurses (0.1 full-time equivalent [FTE]), clinical microbiologists (0.2 FTE), pharmacists (1.0 FTE), and physicians without ID certification (0.1 FTE). The term \"pharmacist-led\" in our context refers to the fact that a dedicated ASP pharmacist, who had 3 years of clinical experience when the ASP was initiated in 2018 and consistently served in this role throughout the study period, functioned as the primary coordinator of the ASP activities. In addition to this dedicated ASP pharmacist, a dedicated drug information pharmacist with 15 years of clinical experience who was also a member of the infection control team participated in the weekly conferences. There were no changes in the ASP team membership or significant FTE variations throughout the study period.\u003c/p\u003e \u003cp\u003eThe intervention method employed was Prospective Audit and Feedback. The ASP team conducted weekly 2-hour conferences to review patients receiving antipseudomonal antibiotics, anti-methicillin-resistant \u003cem\u003eStaphylococcus aureus\u003c/em\u003e drugs, prolonged antimicrobial therapy (\u0026ge;\u0026thinsp;14 days), or those with bloodstream infections. Prior to these conferences, the ASP pharmacist was responsible for identifying patients meeting intervention criteria and preparing comprehensive patient summaries. During these conferences, the team discussed the need for adjustments to treatment duration, additional laboratory tests, changes in antimicrobial agents, and optimization of dosing regimens. The ASP recommendations were communicated to attending physicians through electronic medical records and shared with ward pharmacists via email, with direct communication for urgent cases. The ward pharmacists then collaborated with attending physicians to optimize antimicrobial therapy based on these recommendations. Although an Infection Control Team had been active prior to April 2018, these ASP-related activities had not been previously conducted.\u003c/p\u003e\n\u003ch3\u003eStudy data\u003c/h3\u003e\n\u003cp\u003eThe study period (April 2014 to March 2022) was divided into two periods using April 2018 (when the ASP was implemented) as the cut-off point\u0026mdash;the pre-ASP period (April 2014 to March 2018) and the post-ASP period (April 2018 to March 2022). Clinical records documented by ward pharmacists were extracted from the hospital's electronic medical record system (provided by Software Service, Inc.). Records containing the term \"antimicrobial\" (including alternative terms such as \"antibiotic,\" \"antibacterial,\" and \"anti-infective\") were included for analysis. Records documented by ASP pharmacists were categorized as \"Team records\" in our electronic medical record system and were excluded. Although no standardized templates were used among pharmacists during the study period, we cannot exclude the possibility that individual pharmacists may have used consistent phrasing in their documentation. We also collected data on the number of ward pharmacists and their years of clinical experience during both periods to assess potential confounding factors.\u003c/p\u003e\n\u003ch3\u003eData collection and anonymization\u003c/h3\u003e\n\u003cp\u003eThe data extracted from electronic medical records underwent a two-stage anonymization process. First, the medical information management department at Hirakata Kohsai Hospital replaced identifiable information (such as patient identifiers and medical record numbers) with anonymized codes. Second, the researchers transferred this data to a standalone environment and performed morphological analysis and part-of-speech filtering to detect and remove personal identifiers (such as organization names, personal names, and geographical locations) from the free-text portions. The dataset generated after this two-stage anonymization process was stored in a password-protected Microsoft OneDrive. The original data and anonymization correspondence table were securely managed within the hospital.\u003c/p\u003e\n\u003ch3\u003eText preprocessing\u003c/h3\u003e\n\u003cp\u003eAll analyses were performed using Python 3.9.4. We performed morphological analysis using mecab-python3 (version 1.0.10) with the ipadic dictionary and part-of-speech filtering using the unicodedata2 package (version 16.0.0) on the anonymized dataset. The morphological analysis was executed in the -Ochasen mode for detailed part-of-speech tagging. Parts of speech were classified into nouns, verbs, adjectives, adverbs, interjections, and prenominal adjectives. To extract meaningful words, we excluded certain word categories\u0026mdash;suffixes, numbers, pronouns, auxiliary words, and formal nouns from nouns; auxiliary adjectives from adjectives; and all symbols, particles, and auxiliary verbs.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStandardization of technical terms\u003c/h2\u003e \u003cp\u003eCompound terms were extracted using the termextract package (version 0.12b) in Python. Terms were selected based on their importance scores calculated using the log-frequency ratio method (lr_mode\u0026thinsp;=\u0026thinsp;1, average_rate\u0026thinsp;=\u0026thinsp;True). Words with scores in the top 5% were classified as compound words.\u003c/p\u003e \u003cp\u003eTo standardize terminology variations, we used the Medical Subject Headings (MeSH) and Systematized Nomenclature of Medicine (SNOMED) controlled vocabularies. Terms were standardized according to the following procedure: Two pharmacists with dedicated ASP experience collaboratively searched for equivalent MeSH terms for each extracted term using the MeSH Browser (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://meshb.nlm.nih.gov/\u003c/span\u003e\u003cspan address=\"https://meshb.nlm.nih.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). When an appropriate MeSH term was found, the extracted term was replaced with the standardized MeSH term. If no suitable MeSH term was available, the pharmacists searched for an equivalent SNOMED term using the SNOMED CT Browser (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://browser.ihtsdotools.org/\u003c/span\u003e\u003cspan address=\"https://browser.ihtsdotools.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Terms that were mapped to the same MeSH or SNOMED identifier (e.g., D000900 and 404684003, respectively) were treated as synonyms to ensure consistent terminology. In cases where the pharmacists encountered difficulty in term mapping, consensus was reached through discussion.\u003c/p\u003e \u003cp\u003eSingle alphabet characters and units were excluded as meaningful analysis targets. The dataset with these defined compound words, synonyms, and exclusion words was then reprocessed using morphological analysis to obtain the final processed dataset.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCo-occurrence analysis\u003c/h3\u003e\n\u003cp\u003eWe constructed a co-occurrence network centered on the term \"antimicrobial\" using the networkx package (version 3.2.1) in Python. As \"antimicrobial\" appeared in all records and was expected to show large differences in detection frequency with other terms, we used the Dice coefficient to evaluate the strength of co-occurrence relationships. Considering that the Dice coefficient typically yields higher values than the Jaccard coefficient, we set a conservative threshold of \u0026ge;\u0026thinsp;0.3 to define the presence of co-occurrence relationships. The network was exported in GraphML format and visualized using gephi (version 0.10.1). To ensure network readability, we applied the ForceAtlas layout algorithm, adjusted edge thickness based on the Dice coefficient values, and adjusted node sizes according to their frequency of occurrence. We have depicted only the minimum spanning tree, prioritizing edges with higher coefficients.\u003c/p\u003e \u003cp\u003eCommunity detection within the co-occurrence network was performed using the Louvain algorithm implemented in the python-louvain package (version 0.16). The best partition was determined using modularity optimization (community.best_partition(G)), which identifies groups of densely connected terms. For the interpretation of the communities, two pharmacists with dedicated ASP experience independently reviewed the terms within each community and assigned names reflecting the primary focus of the grouped terms. Any disagreements were resolved through discussion with a third researcher (a faculty member with experience in hospital infection control). The detailed naming process and term lists for each community are provided in the supplementary materials.\u003c/p\u003e\n\u003ch3\u003eCluster analysis\u003c/h3\u003e\n\u003cp\u003eHierarchical clustering was performed using Ward's method in the scipy package (version 1.13.1) on terms detected as nodes in the co-occurrence network. Dissimilarity was initially defined as 1- Dice coefficient, but the final cluster distances were computed using Ward's method, which minimizes the variance increase at each merging step. Based on preliminary examinations of the dendrogram structure, the number of clusters was set to 10% of the total detected terms to balance interpretability. The naming process for the clusters followed the same procedure as described in the co-occurrence analysis section. To evaluate changes in documentation content before and after ASP implementation, we performed cross-tabulation comparing the frequency of occurrence between the pre-ASP and post-ASP periods for the obtained clusters. To exclude the potential impact of the coronavirus disease 2019 (COVID-19) pandemic on our findings, we conducted a sensitivity analysis that excluded data from April 2020 onward (the beginning of the pandemic's impact in Japan). Odds ratios and their 95% confidence intervals were calculated; statistical comparisons between the periods were conducted using Fisher's exact test. Statistical significance was set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eData management\u003c/h2\u003e \u003cp\u003eTo ensure reproducibility, the data extraction conditions, and analysis programs were documented in the electronic medical record system. As specified in the data management policy established by the research ethics committee, the original non-anonymized data were designated to be retained at Hirakata Kohsai Hospital for 3 years from the study completion date, and the processed dataset at Kyoto Pharmaceutical University for 5 years. The policy also specified that all data must be destroyed using methods that prevent restoration after these retention periods.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec13\"\u003e\n \u003ch2\u003eBasic characteristics of the text data\u003c/h2\u003e\n \u003cp\u003eThe basic characteristics of the study data are presented in Table 1. Ward pharmacist staffing remained relatively stable throughout the study period, with a slight increase in the annual average number after ASP implementation (12.8 to 14.5) and consistent average years of clinical experience (7.3 to 7.4 years). The proportion of records containing the term \"antimicrobial\" increased substantially, from 1,353 (3.12%) in the pre-ASP period to 5,155 (7.81%) in the post-ASP period. The preprocessing identified 29 compound words, 20 sets of synonyms, and 37 exclusion words (Additional file 1, Supplementary Tables\u0026nbsp;1 and 2). The morphological analysis resulted in the extraction of 75,148 words (55.5 words per record) in the pre-ASP period and 338,134 words (65.6 words per record) in the post-ASP period.\u003c/p\u003e\n \u003cdiv\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 1\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eBasic characteristics of the study data\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"3\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePre-ASP\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePost-ASP\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWard pharmacist staffing\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNumber of ward pharmacists at period start\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAnnual average number of ward pharmacists\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12.8 ± 2.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e14.5 ± 1.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNew pharmacists during the period\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDeparting pharmacists during the period\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAverage years of clinical experience\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7.3 ± 7.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7.4 ± 5.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDocumentation characteristics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTotal clinical records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e43,374\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e65,982\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRecords containing \"antimicrobial\"\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,353 (3.12%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5,155 (7.81%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAverage characters per record\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e267.1 ± 116.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e353.1 ± 209.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAverage raw terms per record \u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e149.1 ± 67.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e197.1 ± 120.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAverage analyzed terms per record \u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e55.5 ± 24.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e65.6 ± 37.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eValues are presented as mean ± standard deviation where applicable.\u003c/p\u003e\n \u003cp\u003e\u003csup\u003ea\u003c/sup\u003e Raw terms indicate terms extracted after morphological analysis before filtering.\u003c/p\u003e\n \u003cp\u003e\u003csup\u003eb\u003c/sup\u003e Analyzed terms indicate terms used for final analysis after morphological analysis, part-of-speech filtering, and standardization.\u003c/p\u003e\n \u003cp\u003eAbbreviations: ASP, antimicrobial stewardship program.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eTerms co-occurring with \"antimicrobial\"\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eTerms with direct co-occurrence relationships (Dice coefficient ≥0.3) with \"antimicrobial\" were identified—27 terms in the pre-ASP period and 31 terms in the post-ASP period (Table 2). Three terms showed strong co-occurrence relationships (Dice coefficient ≥0.5) in both periods—\"Laboratory Test\" (Dice coefficient: pre-ASP, 0.525; post-ASP, 0.672), \"Creatinine\" (pre-ASP, 0.507; post-ASP, 0.609), and \"Verification\" (pre-ASP, 0.681; post-ASP, 0.584). In the pre-ASP period, \"Necessity\" and \"Particularly\" were characteristically detected, while in the post-ASP period, new strong relationships emerged with \"C-Reactive Protein (CRP)\", \"estimated Glomerular Filtration Rate (eGFR)\", \"White Blood Cell Count (WBC)\", and \"Aspartate Aminotransferase (AST)\".\u003c/p\u003e\n \u003cdiv\u003e\n \u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 2\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eTerms with strong co-occurrence relationships with \"antimicrobials\" (Dice coefficient ≥ 0.3)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"5\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003ePre-ASP (n = 1,353)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003ePost-ASP (n = 5,155)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTerm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFrequency \u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCo-occurrence \u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFrequency \u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCo-occurrence \u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eALT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,677 (32.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.491\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAST\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,740 (33.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.505\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAdministration\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e282 (20.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.345\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,378 (26.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.422\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBacterial Test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e936 (18.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.307\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e338 (25.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.400\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2,457 (47.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.646\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCaution\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e420 (31.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.474\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,114 (21.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.355\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCeftriaxone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e975 (18.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.318\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eContinuation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,041 (20.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.336\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCreatinine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e459 (33.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.507\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2,258 (43.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.609\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDecrease\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,056 (20.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.340\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDiscontinuation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e946 (18.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.310\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDosage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e296 (21.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.359\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDrugs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e428 (31.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.481\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,098 (21.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.351\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eeGFR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e366 (27.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.426\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2,179 (42.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.594\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eExplanation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e378 (27.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.437\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFever\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,044 (20.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.337\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGuidance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e364 (26.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.424\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHospitalization\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,014 (19.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.329\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eInitiation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e358 (26.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.419\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,270 (24.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.395\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLaboratory Test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e482 (35.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.525\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2,609 (50.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.672\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eManagement Method\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e281 (20.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.344\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMed. Adherence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e353 (26.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.414\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMed. Reconciliation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e310 (22.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.373\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e916 (17.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.302\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMed. Taking\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e276 (20.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.339\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e931 (18.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.306\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMedicine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e370 (27.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.430\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eModification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e315 (23.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.378\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,424 (27.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.433\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNecessity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e618 (45.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.627\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,025 (19.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.332\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNeutrophil\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,543 (29.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.461\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eOral Administration\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e327 (24.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.389\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,344 (26.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.414\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eParticularly\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e497 (36.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.537\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePrescription\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e384 (28.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.442\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,258 (24.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.392\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQualitative Test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,497 (29.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.450\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRenal Function\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e346 (25.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.407\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eReport\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e266 (19.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.329\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSBT/ABPC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,062 (20.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.342\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSafety\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e434 (32.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.486\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTablet\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e387 (28.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.445\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,260 (24.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.393\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTake\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e333 (24.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.395\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e979 (19.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.319\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTend\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,158 (22.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.367\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVerification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e698 (51.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.681\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2,124 (41.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.584\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2,113 (41.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.582\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003e\u003csup\u003ea\u003c/sup\u003e Numbers in parentheses indicate the percentage of records containing the corresponding term\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003e\u003csup\u003eb\u003c/sup\u003e Dice coefficients indicate the strength of co-occurrence relationships\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003eBlank cells indicate that the term was not detected or did not meet the threshold for co-occurrence (Dice coefficient ≥ 0.3) in that period.\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003eAbbreviations: ALT, alanine aminotransferase; ASP, antimicrobial stewardship program; AST, aspartate aminotransferase; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; Med., medication; SBT/ABPC, sulbactam/ampicillin; WBC, white blood cell count\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cstrong\u003eNetwork analysis of co-occurring terms\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eA total of 84 co-occurring terms formed a network, with 66 nodes in the pre-ASP period and 63 nodes in the post-ASP period (Figure 1). Louvain algorithm analysis detected seven communities in each period (Additional file 1, Supplementary Table 3). Four communities were common to both periods: Laboratory Monitoring, Antimicrobial Management, Information Collection, and Dose Optimization. The pre-ASP period was characterized by communities related to Medication Review, Prescription Verification, and Patient Education, while the post-ASP period showed newly emerged communities related to Infection Assessment, Microbiological Review, and Severe Infection Management.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eCluster analysis and cross-tabulation\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eWard’s method of hierarchical cluster analysis of the 84 terms identified in the network analysis yielded seven groups (Figure 2). Cross-tabulation analysis comparing the pre- and post-ASP periods showed significant increases in Laboratory Monitoring and Infection Assessment and significant decreases for Prescription Verification, Medication Review, Patient Education, and Information Collection. Only Antimicrobial Management remained consistently high in both periods (Table 3). Sensitivity analysis comparing the pre-ASP period with only the early post-ASP period (April 2018 to March 2020, n=1,400) showed similar patterns to the main analysis, with significant increases in Laboratory Monitoring (OR: 1.79, 95% CI: 1.54–2.09, p\u0026lt;0.01) and Infection Assessment (OR: 2.05, 95% CI: 1.75–2.39, p\u0026lt;0.01) (Additional file 1, Supplementary Table 4).\u003c/p\u003e\n \u003cdiv\u003e\n \u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 3\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eChanges in documentation patterns before and after ASP implementation\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"5\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePre-ASP\u003c/p\u003e\n \u003cp\u003e(n = 1,353)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePost-ASP\u003c/p\u003e\n \u003cp\u003e(n = 5,155)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eOdds ratio (95% CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAntimicrobial Management\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1,337 (98.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5,068 (98.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.70 (0.38 − 1.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLaboratory Monitoring\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e617 (45.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2,934 (56.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.58 (1.39 − 1.78)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt; 0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eInfection Assessment\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e598 (44.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3,214 (62.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.09 (1.85 − 2.36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt; 0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eInformation Collection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e494 (36.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1,405 (27.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.65 (0.57 − 0.74)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt; 0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePrescription Verification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e746 (55.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1,896 (36.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.47 (0.42 − 0.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt; 0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMedication Review\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e992 (73.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3,264 (63.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.63 (0.55 − 0.72)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt; 0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePatient Education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e591 (43.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e917 (17.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.28 (0.24 − 0.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt; 0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003eValues are expressed as n (%). Statistical comparisons were performed using Fisher's exact test. p values \u0026lt; 0.05 were considered statistically significant.\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003eAbbreviations: ASP, antimicrobial stewardship program; CI, confidence interval\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study used text mining analysis to evaluate changes in ward pharmacist interventions for antimicrobial therapy before and after the implementation of an ASP. The analysis revealed characteristic changes in documentation patterns\u0026mdash;specifically, an increase in ID-related terms and a relative decrease in medication guidance-related terms. Network analysis demonstrated a shift from communities related to Medication Review, Prescription Verification, and Patient Education in the pre-ASP period to those focused on Microbiological Review, Infection Assessment, and Severe Infection Management in the post-ASP period.\u003c/p\u003e\n\u003cp\u003eIn the context of antimicrobial therapy, CRP level and WBC are essential indicators for assessing infection severity and treatment efficacy [15]; moreover, renal function and liver function test results (eGFR and AST level, respectively) are crucial for dose adjustment [16, 17], and microbiological test results are essential for appropriate antimicrobial selection and improving treatment outcomes [18, 19]. In our study, new co-occurrence relationships involving these specific laboratory parameters emerged in the post-ASP period, beyond the general association with Laboratory Tests in the pre-ASP period. Additionally, the emergence of the Microbiological Review community suggests that ward pharmacists may have evolved their practices to include more specialized therapeutic support, particularly in terms of pathogen-based antimicrobial selection.\u003c/p\u003e\n\u003cp\u003eThe basic roles of ward pharmacists include medication management activities such as safety verification and providing medication guidance [7, 20, 21]. In our analysis, while terms related to Microbiological Review and Infection Assessment increased after ASP implementation, the proportion of terms related to Medication Review, Patient Education, and Prescription Verification decreased. However, the overall clinical documentation by pharmacists also increased 1.5-fold after ASP implementation, with antimicrobial-related records showing a particularly marked increase (from 3.12% to 7.81% of the total records). These findings suggest that ward pharmacists expanded their expertise into antimicrobial stewardship while maintaining their basic medication management activities.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePrevious studies evaluating ASP effectiveness focused primarily on large hospitals with adequate staff, including one to two ID physicians, at least one dedicated pharmacist, and one or more infection control nurses [10, 22, 23]. In contrast, even among legally designated medical institutions for IDs in Japan, only 33.3% have ID specialists [5], and many hospitals struggle to allocate dedicated ASP staff owing to resource constraints [24]. Although several reports have described the feasibility and effectiveness of ASPs in resource-limited settings [25\u0026ndash;27], very few studies have examined how ward pharmacists adapt their practices to support antimicrobial stewardship in hospitals without ID specialists. Our text mining analysis provides objective evidence on how ward pharmacists can evolve their practice patterns to support antimicrobial stewardship in resource-limited settings.\u003c/p\u003e\n\u003cp\u003eConventional evaluations of ASP activities have primarily focused on quantitative indicators such as antimicrobial use and treatment duration [28]. Although these measures are essential for assessing program effectiveness, they may not fully capture the evolution of ward pharmacist practices and their collaboration with ASP teams. Text mining of medical records has emerged as a complementary analytical approach that can reveal changes in clinical practice patterns [29, 30]. In this study, this analytical method provided unique insights into how ward pharmacists adapt their roles and responsibilities in response to new antimicrobial stewardship initiatives.\u003c/p\u003e\n\u003cp\u003eNonetheless, this study had some limitations. First, text mining cannot be used to establish causal relationships, making it impossible to determine whether the changes in documentation patterns resulted from improved pharmacist expertise or ASP interventions.\u0026nbsp;Although our data showed that the average years of clinical experience of ward pharmacists remained relatively stable throughout the study period (pre-ASP 7.3 years vs. post-ASP 7.4 years), we cannot exclude the possibility that changes in experience levels influenced the results. Additionally, while our sensitivity analysis showed similar results when excluding the COVID-19 pandemic period, we cannot rule out the possibility that major events that significantly impact awareness of IDs, such as the pandemic, may have influenced our findings.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSecond, as this was a single-center retrospective study, the generalizability of our findings remains limited. However, the primary aim of this study was to qualitatively evaluate antimicrobial stewardship activities using a novel text mining approach. We recognize that this investigation serves as a pilot study proposing methodologies for handling unprocessed electronic medical record data containing personal information in text mining research and for evaluating data after morphological analysis. Future studies applying this methodology across healthcare institutions with diverse backgrounds and examining its applicability to other healthcare professionals would contribute to building more robust evidence.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eAs a pilot study, we applied a novel text mining approach to qualitatively evaluate antimicrobial stewardship activities in a single Japanese hospital. Future research applying this methodology across diverse healthcare institutions and examining its applicability to other healthcare professionals will contribute to building more robust evidence for the evolution of pharmacy practice in antimicrobial stewardship.\u003c/p\u003e\n"},{"header":"Abbreviations","content":"\u003cp\u003eADR: adverse drug reaction, ALT: alanine aminotransferase, AMR: antimicrobial resistance, AST: aspartate aminotransferase, ASP: antimicrobial stewardship program, CCr: creatinine clearance, CI: confidence interval, COVID-19: coronavirus disease-2019, CRP: C-reactive protein, eGFR: estimated glomerular filtration rate, FTE: full-time equivalent, ID: infectious disease, Med.: medication, MeSH: Medical Subject Headings, MRSA: methicillin-resistant \u003cem\u003eStaphylococcus aureus\u003c/em\u003e, OR: odds ratio, OTC: over-the-counter, Pharm.: pharmaceutical, SBT/ABPC: sulbactam/ampicillin, SNOMED: Systematized Nomenclature of Medicine, WBC: white blood cell count\u003c/p\u003e\n"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Ethics Committees of Hirakata Kohsai Hospital (approval number: 2022-009) and Kyoto Pharmaceutical University (approval number: E23-029). The study was conducted in accordance with the Declaration of Helsinki and the Ethical Guidelines for Life Science and Medical Research Involving Human Subjects. An opt-out approach was used, where information about the study was made available on the hospital website, and participants were provided the opportunity to refuse participation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analyzed during the current study are not publicly available due to privacy concerns and institutional policies regarding medical records but are available from the corresponding author on reasonable request with approval from the Ethics Committees of Hirakata Kohsai Hospital and Kyoto Pharmaceutical University. The detailed protocol for text preprocessing, including the standardized terminology list and compound word definitions, is available in the supplementary materials.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRI received grants from Kowa Company Ltd., the Japan Pharmaceutical Association, and the Japan Society for the Promotion of Science.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eYM received grants from Pfizer Japan Inc., Kowa Company Ltd., the Japan Pharmaceutical Association, the Japan Society for the Promotion of Science, and the Ministry of Health, Labour and Welfare. YM is also a board member of the Japanese Society of Pharmaceutical Health Care and Sciences and the Japanese Society for Infection Prevention and Control and a committee member of the AMR Clinical Reference Center.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe other authors have no conflicts of interest to declare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKS designed the study protocol, led the antimicrobial stewardship program, collected and analyzed the data, performed the MeSH and SNOMED term mapping, participated in community naming, and drafted the manuscript. SK and RI critically reviewed the manuscript. YM supervised the research project, participated in community naming, and critically reviewed the manuscript. All authors read and approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank the staff of the Infection Control Committee and the Pharmacy Department at the Federation of National Public Service Personnel Mutual Aid Associations Hirakata Kohsai Hospital for managing the ASP. We also thank Hikaru Nishiyama for his contributions to ASP implementation, MeSH and SNOMED term mapping, and community naming. We are grateful to the Medical Information Management Department at the Federation of National Public Service Personnel Mutual Aid Associations Hirakata Kohsai Hospital for their assistance with data extraction and anonymization.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCosgrove SE. The relationship between antimicrobial resistance and patient outcomes: mortality, length of hospital stay, and health care costs. \u003cem\u003eClin Infect Dis\u003c/em\u003e. 2006;42 Suppl 2:S82-9. doi: 10.1086/499406.\u003c/li\u003e\n\u003cli\u003eDoron S, Davidson LE. Antimicrobial stewardship. \u003cem\u003eMayo Clin \u003c/em\u003e\u003cem\u003eProc\u003c/em\u003e. 2011;86(11):1113-23. doi: 10.4065/mcp.2011.0358.\u003c/li\u003e\n\u003cli\u003eBarlam TF, Cosgrove SE, Abbo LM, MacDougall C, Schuetz AN, Septimus EJ, et al. Implementing an antibiotic stewardship program: guidelines by the infectious diseases society of America and the society for healthcare epidemiology of America. \u003cem\u003eClin Infect Dis\u003c/em\u003e. 2016;62(10):e51-77. doi: 10.1093/cid/ciw118.\u003c/li\u003e\n\u003cli\u003eHussaini SMQ. Dearth of infectious diseases physicians as the USA faces a global pandemic. \u003cem\u003eLancet Infect Dis\u003c/em\u003e. 2020;20(6):648-9. doi: 10.1016/S1473-3099(20)30377-7.\u003c/li\u003e\n\u003cli\u003eHagiya H. Shortage and unequal distribution of infectious disease specialists in Japan: how can we refine the current situation? \u003cem\u003ePLOS One\u003c/em\u003e. 2023;18(10):e0291677. doi: 10.1371/journal.pone.0291677.\u003c/li\u003e\n\u003cli\u003eChen K-L, Hunag C-F, Sheng W-H, Chen Y-K, Wang C-C, Shen L-J. Impact of integrated medication management program on medication errors in a medical center: an interrupted time series study. \u003cem\u003eBMC Health Serv Res\u003c/em\u003e. 2022;22(1):796. doi: 10.1186/s12913-022-08178-w.\u003c/li\u003e\n\u003cli\u003eIqbal MJ, Mohammad Ishaq G, Assiri AA. Connecting pharmacists and other health care providers (HCPs) towards drug therapy optimization: A pharmaceutical care approach. \u003cem\u003eInt J Clin Pract\u003c/em\u003e. 2023;2023:3336736. doi: 10.1155/2023/3336736.\u003c/li\u003e\n\u003cli\u003eOda K, Saito H, Jono H. Bayesian prediction-based individualized dosing of anti-methicillin-resistant Staphylococcus aureus treatment: recent advancements and prospects in therapeutic drug monitoring. \u003cem\u003ePharmacol Ther\u003c/em\u003e. 2023;246:108433. doi: 10.1016/j.pharmthera.2023.108433.\u003c/li\u003e\n\u003cli\u003eSukkha S, Rattanavipanon W, Chamroenwit B, Sanganurak M, Nathisuwan S, Chaikledkaew U, Kongwatcharapong J. Quality assessment and cost saving of renal dosing recommendation by clinical pharmacists at medical wards in Thailand. \u003cem\u003eInt J Clin Pharm\u003c/em\u003e. 2020;42(2):610-6. doi: 10.1007/s11096-020-01016-1.\u003c/li\u003e\n\u003cli\u003eOkada N, Azuma M, Tsujinaka K, Abe A, Takahashi M, Yano Y, et al. Clinical impact of a pharmacist-driven prospective audit with intervention and feedback on the treatment of patients with bloodstream infection. \u003cem\u003eAntibiotics (Basel)\u003c/em\u003e. 2022;11(9):1144. doi: 10.3390/antibiotics11091144.\u003c/li\u003e\n\u003cli\u003eNampoothiri V, Hisham M, Mbamalu O, Mohamed ZU, Singh SK, Charani E. Evolution of pharmacist roles in antimicrobial stewardship: A 20-year systematic review. \u003cem\u003eInt J Infect Dis\u003c/em\u003e. 2025;151:107306. doi: 10.1016/j.ijid.2024.107306.\u003c/li\u003e\n\u003cli\u003ePercha B. Modern clinical text mining: A guide and review. \u003cem\u003eAnnu Rev Biomed Data Sci\u003c/em\u003e. 2021;4:165-87. doi: 10.1146/annurev-biodatasci-030421-030931.\u003c/li\u003e\n\u003cli\u003evan de Burgt BWM, Wasylewicz ATM, Dullemond B, Grouls RJE, Egberts TCG, Bouwman A, Korsten EMM. Combining text mining with clinical decision support in clinical practice: A scoping review. \u003cem\u003eJ Am Med Inform \u003c/em\u003e\u003cem\u003eAssoc\u003c/em\u003e. 2023;30(3):588-603. doi: 10.1093/jamia/ocac240.\u003c/li\u003e\n\u003cli\u003eGrob M, Jenders RA, Rappelsberger A, Adlassnig K-P. Integrated perspectives on clinical decision support: A comparative analysis of knowledge management approaches. \u003cem\u003eStud Health Technol Inform\u003c/em\u003e. 2024;316:1822-6. doi: 10.3233/SHTI240785.\u003c/li\u003e\n\u003cli\u003eMuna AM, Alhameed RA. The role of C-reactive protein and white blood cell count as diagnostic, prognostic, and monitoring markers in bacterial orofacial infections. \u003cem\u003eJ Oral Maxillofac Surg\u003c/em\u003e. 2022;80(3):530-6. doi: 10.1016/j.joms.2021.08.158.\u003c/li\u003e\n\u003cli\u003eKeller F, Schr\u0026ouml;ppel B, Ludwig U. Pharmacokinetic and pharmacodynamic considerations of antimicrobial drug therapy in cancer patients with kidney dysfunction. \u003cem\u003eWorld J Nephrol\u003c/em\u003e. 2015;4(3):330-44. doi: 10.5527/wjn.v4.i3.330.\u003c/li\u003e\n\u003cli\u003eHalilovic J, Heintz BH. Antibiotic dosing in cirrhosis. \u003cem\u003eAm J Health Syst Pharm\u003c/em\u003e. 2014;71(19):1621-34. doi: 10.2146/ajhp140031.\u003c/li\u003e\n\u003cli\u003eHamilton WL, Pires S-M, Lippett S, Gudka V, Cross ELA, Llewelyn MJ. The impact of diagnostic microbiology on de-escalation of antimicrobial therapy in hospitalised adults. \u003cem\u003eBMC Infect Dis\u003c/em\u003e. 2020;20(1):102. doi: 10.1186/s12879-020-4823-4.\u003c/li\u003e\n\u003cli\u003eStelling J, Read JS, Fritch W, O\u0026rsquo;Brien TF, Peters R, Clark A, et al. Surveillance of antimicrobial resistance and evolving microbial populations in Vermont: 2011-2018. \u003cem\u003eExpert Rev Anti Infect Ther\u003c/em\u003e. 2020;18(10):1055-62. doi: 10.1080/14787210.2020.1776114.\u003c/li\u003e\n\u003cli\u003eSalmasi S, Tsao NW, Li K, Shaske JN, Marra CA, Lynd LD. Characterization of pharmacist-based medication management services in a community pharmacy. \u003cem\u003eRes Social Adm Pharm\u003c/em\u003e. 2020;16(2):178-82. doi: 10.1016/j.sapharm.2019.04.051.\u003c/li\u003e\n\u003cli\u003eBainbridge J, Barnhart R, Fuller R, Hellerslia VT, Kidd J, Merrill S, et al. The role of clinical pharmacists in patient-centric comprehensive multiple sclerosis care. \u003cem\u003eInt J MS Care\u003c/em\u003e. 2024;26(1):1-7. doi: 10.7224/1537-2073.2022-051.\u003c/li\u003e\n\u003cli\u003eNathwani D, Varghese D, Stephens J, Ansari W, Martin S, Charbonneau C. Value of hospital antimicrobial stewardship programs [ASPs]: A systematic review. \u003cem\u003eAntimicrob Resist Infect Control\u003c/em\u003e. 2019;8:35. doi: 10.1186/s13756-019-0471-0.\u003c/li\u003e\n\u003cli\u003eGiamarellou H, Galani L, Karavasilis T, Ioannidis K, Karaiskos I. Antimicrobial stewardship in the hospital setting: A narrative review. \u003cem\u003eAntibiotics (Basel)\u003c/em\u003e. 2023;12(10):1557. doi: 10.3390/antibiotics12101557.\u003c/li\u003e\n\u003cli\u003eMaeda M, Muraki Y, Kosaka T, Yamada T, Aoki Y, Kaku M, et al. Essential human resources for antimicrobial stewardship teams in Japan: estimates from a nationwide survey conducted by the Japanese Society of Chemotherapy. \u003cem\u003eJ Infect Chemother\u003c/em\u003e. 2019;25(9):653-6. doi: 10.1016/j.jiac.2019.05.012.\u003c/li\u003e\n\u003cli\u003eBanerjee S, Gupta N, Ray Y, Kodan P, Khot WY, Fazal F, et al. Impact of trainee-driven Antimicrobial Stewardship Program in a high burden resource-limited setting. \u003cem\u003eInfez Med\u003c/em\u003e. 2020;28(3):367-72. https://\u003cu\u003epubmed.ncbi.nlm.nih.gov/\u003c/u\u003e32920572.\u003c/li\u003e\n\u003cli\u003eSawada K, Inose R, Goto R, Nakatani T, Kono S, Muraki Y. Effectiveness of noncertified pharmacist-led antimicrobial stewardship programs in a medium-sized hospital without an infectious disease specialist: A retrospective pre-post study. \u003cem\u003eOpen Forum Infect Dis\u003c/em\u003e. 2023;10(3):ofad116. doi: 10.1093/ofid/ofad116.\u003c/li\u003e\n\u003cli\u003eNakano Y, Hirai T, Murata M, Yasukochi H, Ura K, Sueyasu Y, et al. Impact of pharmacist-driven antimicrobial stewardship interventions in a secondary care facility in Japan: A pragmatic quasi-experimental study. \u003cem\u003eJ Infect Chemother\u003c/em\u003e. 2025;31(1):102503. doi: 10.1016/j.jiac.2024.08.018.\u003c/li\u003e\n\u003cli\u003eSeptimus E. Antimicrobial stewardship-qualitative and quantitative outcomes: the role of measurement. \u003cem\u003eCurr Infect Dis Rep\u003c/em\u003e. 2014;16(11):433. doi: 10.1007/s11908-014-0433-x.\u003c/li\u003e\n\u003cli\u003eMarafino BJ, Park M, Davies JM, Thombley R, Luft HS, Sing DC, et al. Validation of prediction models for critical care outcomes using natural language processing of electronic health record data. \u003cem\u003eJAMA Netw Open\u003c/em\u003e. 2018;1(8):e185097. doi: 10.1001/jamanetworkopen.2018.5097.\u003c/li\u003e\n\u003cli\u003eWieland-Jorna Y, van Kooten D, Verheij RA, de Man Y, Francke AL, Oosterveld-Vlug MG. Natural language processing systems for extracting information from electronic health records about activities of daily living. A systematic review. \u003cem\u003eJAMIA Open\u003c/em\u003e. 2024;7(2):ooae044. doi: 10.1093/jamiaopen/ooae044.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"antimicrobial stewardship program, ward pharmacist, text mining, co-occurrence network analysis, medical records, infectious disease, medium-sized hospital","lastPublishedDoi":"10.21203/rs.3.rs-6199166/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6199166/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eAntimicrobial stewardship programs (ASPs) are essential for optimizing antimicrobial use, but many medium-sized hospitals lack infectious disease (ID) specialists. Ward pharmacists can contribute to ASPs, but the qualitative changes in their practice patterns after ASP implementation remains unclear. We aimed to explore the potential of text mining as a novel methodology to evaluate changes in ward pharmacist antimicrobial management practices after ASP implementation in a medium-sized hospital without ID physicians.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe conducted a retrospective observational analysis of data documented in clinical records by ward pharmacists in a 313-bed community hospital from April 2014 to March 2022. The ASP team conducted weekly reviews of targeted patients, provided feedback to physicians, and shared recommendations with ward pharmacists who then collaborated to optimize antimicrobial therapy. Using Python-based text mining with standardized technical terms and compound word extraction, we performed morphological analysis, co-occurrence network analysis, and hierarchical clustering to evaluate documentation patterns before and after ASP implementation in April 2018. Co-occurrence relationships were assessed using Dice coefficients (threshold, \u0026ge;\u0026thinsp;0.3), and communities were detected using the Louvain algorithm. Changes in documentation patterns were compared using Fisher's exact test.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe analysis included 1,353 pre-ASP and 5,155 post-ASP clinical records containing antimicrobial-related terms, which increased from 3.12\u0026ndash;7.81% of the total pharmacy records. New strong co-occurrence relationships emerged in the post-ASP period for several laboratory parameters (c-reactive protein, 0.646; estimated glomerular filtration rate, 0.594; and white blood cell count, 0.582). Network analysis revealed a shift from medication-focused communities (Medication Review, Prescription Verification, and Patient Education) to infection-focused communities (Infection Assessment, Microbiological Review, and Severe Infection Management). Although Antimicrobial Management was consistently used in both periods (odds ratio [OR]: 0.70, 95% confidence interval [CI]: 0.38\u0026ndash;1.20), cross-tabulation analysis increased significantly in Laboratory Monitoring (OR: 1.58, 95% CI: 1.39\u0026ndash;1.78) and Infection Assessment (OR: 2.09, 95% CI: 1.85\u0026ndash;2.36).\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThis pilot application of text mining demonstrated potential as a novel methodology for objectively evaluating qualitative changes in clinical practice patterns following ASP implementation, successfully identifying shifts in pharmacists' documentation focus and providing a foundation for future multi-center validation studies across diverse healthcare settings.\u003c/p\u003e","manuscriptTitle":"Qualitative Changes in Clinical Records After Implementation of Pharmacist-Led Antimicrobial Stewardship Program: A Text Mining Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-04 11:17:40","doi":"10.21203/rs.3.rs-6199166/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"15cc0b2e-f645-47d4-b27c-b4e53e868ecf","owner":[],"postedDate":"April 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-04-28T16:01:39+00:00","versionOfRecord":{"articleIdentity":"rs-6199166","link":"https://doi.org/10.1186/s40780-025-00439-0","journal":{"identity":"journal-of-pharmaceutical-health-care-and-sciences","isVorOnly":false,"title":"Journal of Pharmaceutical Health Care and Sciences"},"publishedOn":"2025-04-23 15:57:31","publishedOnDateReadable":"April 23rd, 2025"},"versionCreatedAt":"2025-04-04 11:17:40","video":"","vorDoi":"10.1186/s40780-025-00439-0","vorDoiUrl":"https://doi.org/10.1186/s40780-025-00439-0","workflowStages":[]},"version":"v1","identity":"rs-6199166","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6199166","identity":"rs-6199166","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}