• KSAN
  • Contact us
  • E-Submission

Korean J Adult Nurs : Korean Journal of Adult Nursing

OPEN ACCESS
mobile search button mobile menu button
Search
Advanced Search
ABOUT
BROWSE ARTICLES
EDITORIAL POLICY
FOR CONTRIBUTORS

Articles

Page Path

Invited Article

Machine Learning Applications in Nursing-Affiliated Research: A Systematic Review

Eun Joo Kim1orcid, Seong Kwang Kim2orcid
Korean Journal of Adult Nursing 2025;37(3):189-214.
Published online: August 29, 2025

1Associate Professor, Department of Nursing, Gangneung-Wonju National University, Wonju, Korea

2Ph.D. Student, Department of Nursing, Gangneung-Wonju National University, Wonju, Korea

Corresponding author: Seong Kwang Kim Department of Nursing, Gangneung-Wonju National University, 150 Namwon-ro, Heungop-myeon, Wonju 26403, Korea. Tel: +82-33-760-8650 Fax: +82-33-760-8641 E-mail: ksk1677@naver.com
• Received: March 27, 2025   • Revised: May 22, 2025   • Accepted: July 30, 2025

© 2025 Korean Society of Adult Nursing

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc-nd/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

  • 869 Views
  • 25 Download
next

Full Article

Download PDF Download PDF
  • Purpose
    This study analyzed the methodological characteristics of machine learning (ML) applications in nursing research, evaluated their reporting quality against standardized guidelines, and assessed progress toward clinical implementation.
  • Methods
    A PRISMA-compliant systematic review (PROSPERO CRD42024595877) searched nine English- and Korean-language databases through September 27, 2024. Included studies applied ML to a nursing question and had at least one nursing-affiliated author. Two reviewers independently extracted data following the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework. Reporting quality was appraised using the TRIPOD+AI checklist.
  • Results
    Of 125 included studies, supervised learning predominated (93.6%), with random forest, logistic regression, and support vector machines as common algorithms. The most frequent performance metrics were the area under the receiver operating curve and accuracy. Mean TRIPOD+AI compliance was 50.4% (standard deviation=9.37), with reporting quality lowest for data preparation (48.0%) and class imbalance handling (22.4%). Research focused on predicting pressure injuries, falls, and readmissions. Only seven studies described clinical deployment, often citing ethical or workflow barriers.
  • Conclusion
    While ML studies in nursing are increasing and show strong discriminatory accuracy, their impact is limited by inconsistent reporting, limited external validation, and rare clinical deployment. Translating these algorithms into practice requires adopting comprehensive reporting guidelines like TRIPOD+AI, documenting each CRISP-DM phase, and integrating nurse-centered decision-support pathways.
  • Key Words: Nursing; Machine learning; Systematic review
1. Background
Artificial intelligence (AI) encompasses technologies that enable computers and machines to mimic human capabilities such as learning, understanding, problem-solving, decision-making, creativity, and autonomy [1]. AI is recognized as a defining technology of the Fourth Industrial Revolution [2], with the field continually achieving remarkable milestones—from Google DeepMind’s AlphaGo in 2016 to OpenAI’s ChatGPT in 2022 [3].
Machine learning (ML), a branch of AI, focuses on training algorithms to develop predictive or classification models based on data, allowing for learning and inference without explicit programming [1]. Enhanced computing power and the availability of large-scale datasets have fueled the rapid advancement and widespread adoption of ML across diverse sectors, including healthcare, finance, manufacturing, and transportation [4]. Due to its exceptional predictive capabilities, interdisciplinary research involving ML has expanded significantly, establishing it as a key technology for tackling contemporary challenges [4,5].
ML applications in nursing research are also becoming more prevalent [6]. For example, prior studies have predicted nursing students’ graduation likelihood using academic performance in their first year, achieving over 80% accuracy, which increased to as high as 99% with three-year longitudinal data, thus enabling automated, personalized assessments for students at risk of attrition [7]. ML has also been used to predict the 30-day readmission probability for heart failure patients based on multiple clinical variables [8], and to forecast nurse turnover rates using personnel data [9]. In addition, automated extraction and analysis of nursing documentation have enhanced administrative decision-making, improving both speed and accuracy [10]. Successful ML implementation allows nurses to dedicate more time to direct patient care, thereby raising the overall quality of nursing services [11].
Nevertheless, several limitations remain regarding the application of ML in nursing research. When predicting rare events—such as specific disease occurrences or serious medical conditions—datasets are often imbalanced, containing significantly fewer event cases than non-events. Such an imbalance can undermine a model’s ability to accurately detect rare cases, compromising both predictive performance and generalizability [12]. Moreover, heterogeneity in data collection methods—including varying formats, measurement techniques, and terminologies—leads to non-standardized datasets [12]. This heterogeneity complicates data preprocessing, diminishes model consistency and reliability [9], and is further aggravated by the challenges of data collection and limited participant recruitment commonly encountered in nursing studies, often resulting in small sample sizes [9]. Collectively, these factors adversely affect the performance and clinical relevance of ML models.
These constraints may ultimately impair ML models’ performance and clinical validity. To harness the potential of ML in nursing, systematic analyses of current research are essential. Although the use of ML in nursing studies is expanding, comparison across studies remains difficult due to non-standardized evaluation criteria, inconsistent data handling methods, and variable reporting practices. Furthermore, assessments of practical applicability—including clinical utility, cost-effectiveness, and patient safety—are needed. Accordingly, this systematic review provides a comprehensive evaluation of the current landscape of ML in nursing research. Specifically, this study aims to (1) systematically identify and describe methodological characteristics using the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework; (2) critically assess methodological rigor and reporting transparency according to the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis for Artificial Intelligence (TRIPOD+AI) checklist; and (3) evaluate practical deployment and clinical readiness. Through this multi-dimensional analysis, we aim to identify key gaps and offer evidence-based recommendations to guide future research toward improved reproducibility, standardization, and, ultimately, clinical impact.
2. Conceptual Framework
In this systematic review, the CRISP-DM methodology [13] was adopted as the conceptual framework for systematically evaluating studies employing ML techniques.
Earlier standardized data analysis methodologies, such as Knowledge Discovery in Databases (KDD) and Sample, Explore, Modify, Model, Assess (SEMMA), also include data preprocessing stages. However, these frameworks tend to focus primarily on analytical techniques rather than comprehensively guiding the entire analysis process [14]. In contrast, CRISP-DM offers a structured workflow with explicit guidance for each phase, ensuring alignment with broader business objectives. A key feature of CRISP-DM is its iterative feedback loop, which allows movement between phases as needed. For instance, if problems are identified during data preparation, the process can return to the business understanding phase to revise objectives accordingly. This flexibility sets CRISP-DM apart from more linear models like KDD and SEMMA, making it especially suitable for practice-oriented data analysis and clinical applications in nursing research.
CRISP-DM is a standardized process model for data mining projects, encompassing six phases: (1) business understanding, (2) data understanding, (3) data preparation, (4) modeling, (5) evaluation, and (6) deployment. In this review, study procedures were structured based on the CRISP-DM methodology, as outlined in Table 1. This conceptual framework enabled a systematic evaluation of nursing literature involving ML techniques.
The protocol for this study was registered with PROSPERO (registration number: CRD42024595877) on February 10, 2024.
1. Study Design
This study is a systematic review aimed at identifying and critically appraising the methodological characteristics of ML applications in nursing research. The review was conducted according to the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions 6.4 [15] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [16].
2. Key Questions and Eligibility Criteria
To achieve the study’s overall objective of evaluating the methodological landscape and clinical readiness of nursing ML research, the following key questions were addressed. These were structured to align with the CRISP-DM framework and our central themes of reporting standards and practical application:
In the business understanding phase (phase 1), we examine the research objectives and questions of nursing studies applying ML techniques, and the reasons for applying ML in these studies. In the data understanding phase (phase 2), we investigate the data sources and data quality of nursing studies applying ML techniques. The data preparation phase (phase 3) examines the data preprocessing methods used in nursing studies applying ML techniques and the transparency with which they are reported to ensure reproducibility. In the modeling phase (phase 4), we assess how model training, validation, and testing are performed in nursing studies applying ML techniques. The evaluation phase (phase 5) considers the performance evaluation metrics used in nursing studies applying ML techniques and examines the extent to which they reflect clinical utility beyond algorithmic accuracy. The deployment phase (phase 6) addresses whether ethical considerations have been addressed and the degree to which these models have been practically deployed in clinical settings, bridging the gap from research to practice.
Inclusion criteria were as follows: (1) studies involving nursing research, operationalized as having at least one author affiliated with a nursing-related institution (e.g., nursing school or research center), chosen as an objective and reproducible proxy for identifying research likely to be informed by a nursing perspective; and (2) direct application of ML methodologies (e.g., predictive or classification models, including single-layer Artificial Neural Networks [ANNs]). During data extraction, studies were screened to ensure that only those employing ML techniques were included. Deep learning (DL) studies were identified and excluded based on the use of multi-layer neural networks. Studies were excluded if they: (1) lacked authors with a nursing background or affiliation; (2) did not directly apply ML as a methodology (e.g., studies evaluating ML-based wearable devices without directly applying ML); (3) were not human-subject studies (e.g., animal or plant experiments, or robot development); (4) were review articles; (5) lacked full research results (e.g., abstracts or poster presentations only); or (6) used DL methodologies (e.g., multi-layer neural networks). The exclusion of studies using DL was intentional to ensure methodological coherence. Traditional ML (e.g., random forest, support vector machine [SVM], logistic regression) and DL (e.g., models with multi-layer neural networks such as Convolutional Neutral Networks [CNNs] or Recurrent Neural Networks [RNNs]) often differ significantly regarding data types, feature engineering, and model complexity. Combining these distinct paradigms would have introduced substantial heterogeneity, potentially obscuring trends specific to each. Thus, this review focuses on traditional ML techniques, which represent the foundational and most prevalent approach in the nursing literature to date.
3. Literature Search and Selection Process
Two researchers independently conducted the literature search from September 27 to September 28, 2024, including all literature published up to September 27, 2024. To effectively identify nursing-related literature, electronic databases were selected using the Core, Standard, Ideal (COSI) model proposed by the National Library of Medicine (NLM) [17]. Korean databases included KoreaMed, Kmbase, KISS, NDSL, and KISTI, while international databases comprised Cochrane CENTRAL, MEDLINE, and Embase. Additional sources included PubMed (provided by the NLM), specialized databases such as CINAHL and PsycINFO, and broad academic databases like Scopus and Web of Science.
Advanced searches were conducted using the Participant, Intervention, Comparison, Outcome, Time, Setting, and Study Design (PICOTS-SD) framework. The core search strategy combined terms related to participants (using keywords such as “nurs*” to capture variations of nurse/nursing) with terms representing study design focused on ML methodologies. Specifically, searches included “machine learning” together with techniques such as “classif*,” “regress*,” “predict*,” “forecast*,” “cluster*,” “dimensionality reduction,” “reinforcement learning,” or “policy learning” (using OR logic within this group and AND logic to combine with “nurs*”). The final search string was: ((Nurs*) AND ((Machine Learning) AND (Classif* OR Regress* OR Predict* OR Forecast* OR Cluster* OR "Dimensionality Reduction" OR "Reinforcement Learning" OR "Policy Learning"))).
Keywords were searched within Titles and Abstracts; when simultaneous searching in both fields was not possible, Abstract searches were prioritized. The goal was a comprehensive and systematic review of the extensive nursing-related literature. To ensure inclusivity and capture a broad range of relevant terms, truncation with the asterisk (*) was intentionally applied. For example, searching for “Nurs*” retrieved terms such as “nurses,” “nursing,” “nursing student,” and “nurse aide.” Controlled vocabularies like MeSH (for PubMed) or Emtree (for Embase) were not used. With the exception of database-specific adjustments to the Title/Abstract field, the same search string was otherwise used unchanged across all databases (Supplementary Table 1).
The initial database search yielded 540 articles from PubMed, 611 from Embase, 438 from MEDLINE, 72 from the Cochrane Library, 181 from CINAHL, 52 from PsycINFO, 548 from Web of Science, 1,199 from Scopus, and 12 from ScienceON, totaling 3,653 records. After removing 2,102 duplicates using EndNote 21, 1,551 records remained for screening. Titles and abstracts were reviewed, resulting in the exclusion of 1,285 records that were clearly unrelated to the research topic. Because detailed methodological information is often unavailable at this stage, a conservative approach was taken: records were excluded only when no relevance was evident, while ambiguous or potentially related items were retained for full-text screening. Subsequently, 266 full-text articles were assessed against the exclusion criteria, with 114 meeting the inclusion criteria. An additional 11 studies were identified by screening the reference lists of related systematic reviews. Ultimately, 125 studies were included in the final synthesis (Figure 1). The full list of the 125 included studies is provided in Appendix 1, cited in-text with an “A” prefix (e.g., [A1]). The 152 studies excluded at the full-text screening stage, along with reasons for exclusion, are listed in Appendix 2 and are cited with an “E” prefix (e.g., [E1]) when referenced.
4. Quality Assessment of Included Studies
The quality of the included studies was assessed using the TRIPOD+AI checklist [18]. TRIPOD+AI is an extension of the original TRIPOD 2015 guidelines, providing a comprehensive 27-item assessment tool specifically designed to ensure transparent reporting in ML-based predictive model studies. This checklist is optimized for evaluating both methodological rigor and the reliability of result interpretation in medical AI research. TRIPOD+AI systematically evaluates key domains, including: standardized reporting of model development and validation processes (e.g., data sources [item 5a], justification of sample size [item 10], and handling of missing data [item 11]), thereby emphasizing methodological transparency; AI-specific methodological considerations such as hyperparameter tuning (item 12c), class imbalance handling (item 13), and algorithmic fairness evaluation (item 14); and clinical applicability, including interpretation of model outcomes (item 15) and user interaction requirements in real-world clinical environments (item 27b), thus facilitating evaluation of nursing utility. Two independent reviewers (reviewer A and reviewer B), both trained in the TRIPOD+AI guidelines, assessed each included study. Discrepancies were discussed face-to-face; if consensus was not reached, a third reviewer (reviewer C) adjudicated.
5. Ethical Considerations
This study is a secondary data analysis of previously published literature, conducted using systematic review methodology. Ethical review approval was requested from the Institutional Review Board (IRB) of the researchers’ affiliated institution (GWNUIRB-R2024-65), which confirmed that this research does not involve human subjects and is therefore exempt from further ethical review.
6. Data Analysis
The selected final articles were descriptively summarized in a case report format using Excel 2016 (Microsoft, Redmond, WA, USA). The case report template consisted of 29 items structured according to the phases of the CRISP-DM methodology, as shown in Table 2. To improve coherence, the RESULTS section is organized in a two-tier hierarchy: (1) Methodological Characteristics of ML Studies, sub-organized by the CRISP-DM stages, and (2) Reported Outcomes and Practical Impact.
1. Methodological Characteristics of ML Studies

1) Quality assessment of included studies

The methodological quality of the 125 included studies was evaluated using the 27-item TRIPOD+AI checklist, which assesses transparent reporting in ML-based predictive model research. The average compliance rate was 50.4% (standard deviation=9.37), with a range from 22.9% to 79.2%. Most studies consistently reported on model development or validation (98.4%), data sources (96.8%), and outcome definitions (98.4%). However, compliance was notably lower for data preparation (48.0%), class imbalance handling (22.4%), and algorithmic fairness (2.4%). No studies fully adhered to abstract reporting guidelines or shared their protocols publicly. These reporting gaps highlight persistent challenges in achieving transparency and reproducibility, particularly in areas such as preprocessing and fairness, which are critical issues for nursing-related applications. Only seven studies reported real-world deployment, further underscoring the limited clinical translation of ML models in nursing research. Notably, the lowest quality scores were observed in the studies by Choi et al. [A8] and Chavan et al. [A92], both with a score of 22.9%. In contrast, the highest quality score was achieved by Shao et al. [A106], at 79.2%.

2) General characteristics of the included studies

An analysis of country of origin showed that the United States produced the largest share of studies (n=49, 30.8%), followed by China (n=23, 14.5%) and South Korea (n=17, 10.7%).
Examining the publication years from 2006 to 2024, only a few studies appeared in the early years: 2 (1.6%) in 2006, and just 1 each (0.8%) in 2007 and 2008. A significant increase occurred beginning in 2020, with 11 studies (8.8%) published in 2020, 20 (16.0%) in 2021, 18 (14.4%) in 2022, and 24 (19.2%) in 2023. The highest number, 26 studies (20.8%), was published in 2024.
Analysis of research team composition revealed that studies with 0%–25% nurse involvement were least common (n=16, 12.8%), whereas those with 75%–100% nurse involvement were most frequent (n=42, 33.6%), indicating that nurses comprised at least 25% of the team in the majority of studies.
Among the 125 studies, supervised learning methods were most prevalent (n=117, 93.6%), while unsupervised learning was rare (n=5, 4.0%), and a mixed approach appeared in three studies (2.4%).
In terms of algorithm types, classification algorithms dominated (n=101, 80.8%), followed by regression (n=14, 11.2%) and clustering algorithms (n=7, 5.6%). The most common research objective was pressure injury/ulcer prediction (n=24, 19.2%), followed by readmission- or utilization-related outcomes (n=17, 13.6%), and fall-risk prediction (n=10, 8.0%). The remaining studies (n=74, 59.2%) covered a broad range of topics, including mental health screening, infection detection, workload assessment, and violence prevention.
With respect to journal distribution, CIN: Computers, Informatics, Nursing was the most frequently represented journal (n=7, 5.6%), followed by Applied Clinical Informatics, International Journal of Environmental Research and Public Health, and Journal of the American Medical Informatics Association, each with four studies (3.2%). Additionally, BMC Medical Informatics and Decision Making, International Journal of Medical Informatics, Journal of Advanced Nursing, Journal of Nursing Management, and Nursing Research each published three studies (2.4%) (Table 3).

3) Summary of results according to phase 1: business understanding

Analysis of the 125 included studies identified pressure ulcers, falls, and hospital readmissions as major research foci. Pressure ulcer studies concentrated on early prediction of risk among hospitalized and postoperative patients, aiming to improve nursing quality and patient safety through targeted prevention strategies. Many emphasized the use of ML-based predictive models to identify high-risk patients and enable preventive nursing interventions. Studies related to falls focused on predicting fall risk in both hospitalized patients and nursing home residents, using ML to analyze clinical records, identify key risk factors, and enhance intervention strategies. Research on readmission prediction and management prioritized early identification of high-risk patients to improve management and reduce healthcare costs through timely intervention. These studies often targeted populations such as patients with diabetes, pediatric patients, and individuals requiring post-acute care, developing predictive models based on ML techniques.
Regarding data mining analysis frameworks, the majority of studies (n=100) did not specify an analytic framework (“NI”—no information). Of those that did, CRISP-DM was most commonly mentioned (n=3), followed by KDD (n=2). Other frameworks cited once each included Data, Information, Knowledge, Wisdom (DIKW), the Ahituv Information Flow Model (Ahituv IFM), Plan-Do-Study-Act (PDSA), and the Healthcare Process Modeling to Phenotype Clinician Behaviors Framework (HPM-ExpertSignals). Supplementary Table 2 provides detailed descriptions of the objectives and data mining frameworks employed by all included studies.

4) Summary of results according to phase 2: data understanding

Analysis of the 125 included studies showed that R, Python, and SPSS were the most commonly used tools for data analysis. Additional software packages reported in some studies included SAS (SAS Institute, Cary, NC, USA), Weka (University of Waikato, Hamilton, New Zealand), MATLAB (MathWorks, Natick, MA, USA), MeCab (Nara Institute of Science and Technology, Ikoma, Japan), JMP Pro (SAS Institute), and Modeller (University of California, San Francisco, CA, USA). However, 21 studies did not explicitly specify which software was used. Regarding data sources, electronic medical records and electronic health records (EHRs) were the primary datasets. Other major sources included survey data, student academic records, nursing documentation, and hospital administrative data. Studies utilizing image or audio data were relatively rare.
Exploratory data analysis (EDA) primarily included descriptive statistical methods such as frequency analysis, percentages, mean, and standard deviation. Other methods, such as minimum and maximum values, interquartile range, and data ranges, were less frequently employed. Some studies included natural language processing analyses utilizing text mining, while clustering and dimensionality reduction analyses were occasionally employed. Regarding study populations, patient-centered research was most common, followed by studies targeting nurses, nursing home residents, and nursing students. Studies involving hospital administrators and older adults residing in the community were also present. The study with the largest sample analyzed data from approximately 3.6 million patients, followed by another study analyzing around 1.93 million patient episodes. Another large-scale study included about 1.53 million patients.
In contrast, the smallest study involved around 1,300 qualitative data points collected from 43 patients. Supplementary Table 3 provides detailed information on the software, data sources and types, EDA methods, study populations, and sample sizes for each included study.

5) Summary of results according to phase 3: data preparation

Analysis of data preprocessing techniques across the 125 included studies showed that handling missing data was the most frequently employed approach. Methods for managing missing values included simple or multiple imputation, K-nearest neighbors imputation, omission of missing cases, and various recoding strategies. Standardization and encoding were the next most common preprocessing techniques. Standardization methods included Z-score standardization, scaling, and zero-centering, while encoding approaches comprised label encoding, one-hot encoding, use of dummy variables, and binary recoding. Feature selection methods included variable selection, identification of key predictors, the least absolute shrinkage and selection operator, and recursive feature elimination with cross-validation. Normalization techniques such as min-max normalization and other data normalization strategies were also described.
The majority of studies used fewer than 50 predictor variables, although studies employing text mining techniques (e.g., term frequency-inverse document frequency or Word2Vec), imaging, or sensor data often involved high-dimensional datasets with more than 100 predictors, sometimes ranging from 800,000 to 1,000,000 dimensions. In certain clinical and nursing studies, the number of independent variables varied depending on the feature engineering process and the study stage.
The most common outcome or target variables were related to pressure ulcers, such as pressure injury risk, occurrence, and hospital-acquired pressure injury. Studies on falls often analyzed fall occurrence and severity or type, from binary outcomes (fall/no fall) to more nuanced classifications. Readmission and healthcare utilization studies targeted outcomes such as readmission within defined timeframes (e.g., 30-day, 90-day), emergency department visits, hospital length of stay, and frequency of hospital visits. Additional studies focused on predicting infections (e.g., sepsis, urinary tract infections), mental and psychological states (depression, anxiety, burnout, suicide risk), and mortality.
For data partitioning, simple proportional splits (such as 70% training/30% testing or 80% training/20% testing) were common. Cross-validation methods, including 10-fold, 5-fold, 3-fold, and leave-one-out cross-validation, were also frequently applied. Supplementary Table 4 provides detailed information on data preprocessing methods, predictor and outcome variables, and data partitioning strategies used in the 125 included studies.

6) Summary of results according to phase 4: modeling

Analysis of the 125 included studies revealed that random forest was the most frequently employed ML algorithm, followed by logistic regression (including several modified forms), SVM (including variants), decision trees (including variants), and eXtreme Gradient Boosting (XGBoost). Other algorithms reported included CatBoost, Gradient Boosting Machine (GBM), LightGBM, Elastic Net, stochastic gradient descent, and Bayesian networks. Regarding hyperparameter optimization and parameter settings, most studies either used default parameters or did not report any details (“NI”). This indicates a general lack of explicit reporting or limited use of advanced hyperparameter tuning. Among studies that described tuning approaches, grid search was the most commonly used method. Supplementary Table 5 contains detailed descriptions of the ML models and hyperparameter optimization methods used in each study.
2. Reported Outcomes and Practical Impact

1) Summary of results according to phase 5: evaluation

Analysis of performance metrics and ML model usage across the 125 studies showed that the most frequently reported metric was the area under the receiver operating characteristic curve (AUC-ROC), cited in 68 studies. Accuracy was reported in 64 studies, while sensitivity and specificity appeared in 52 and 47 studies, respectively. Other reported metrics included F1-score (n=26), precision (n=25), positive predictive value (n=20), recall (n=19), and negative predictive value (n=18).
Random forest was most often identified by individual studies as the highest-performing algorithm in their comparisons (35 studies), either as the exclusive model or within comparative analyses. XGBoost was specifically reported as the top performer in 11 studies, either as “XGBoost” or “Only one used (XGBoost),” followed by GBM (n=8), logistic regression (n=15), and SVM-based methods (n=7). Additional algorithms such as M5P tree, ANNs, and Bayesian networks were also used in several studies.
For feature importance, 42 studies provided no information (“NI”) or did not report explicit importance analyses. “Feature importance” was specifically reported in 19 studies, with Shapley additive explanations (SHAP) (n=10), information gain (n=6), permutation importance, recursive feature elimination, and Gini impurity each appearing multiple times. Other methods occasionally used included logistic regression coefficients, mutual information, Markov blanket analysis, and normalized importance. Supplementary Table 6 gives detailed accounts of performance metrics, best-performing models, and feature importance techniques for each included study.

2) Summary of results according to phase 6: deployment

Among the 125 studies analyzed, hospital settings were the most frequently represented research environments, followed by home healthcare, nursing homes, community-based settings, and general healthcare contexts. IRB ethical approval was obtained in 95 studies (76.0%), while 30 studies (24.0%) did not report or obtain IRB approval.
Practical deployment and real-world implementation were described in seven studies. Examples included development of a mobile application for early delirium screening in long-term care, establishment of EHR systems for automated patient risk-factor collection, creation of AI-based systems for assessing emergency department visits and hospitalization risk, web-based applications applying ML models for stroke mortality prediction, pressure injury risk prediction for intensive care unit patients, and provision of online decision-support platforms along with real-time telemedicine services. Supplementary Table 7 provides detailed information about each study’s setting, IRB approval status, and deployment cases.
A summary of all findings based on the CRISP-DM framework is provided in Table 4.
This review was designed to systematically evaluate the methodological characteristics, reporting quality, and clinical translation of ML in nursing research. Our findings reveal a field facing a critical paradox. While the use of ML is rapidly expanding, its translation into robust, reproducible, and clinically impactful tools is hindered by significant methodological shortcomings. This is starkly illustrated by a mean TRIPOD+AI compliance rate of only 50.4%, which signals substantial gaps in both reporting transparency and methodological rigor [18]. Major deficiencies in data preparation (48.0%), class imbalance handling (22.4%), and algorithmic fairness (2.4%) undermine transparency, particularly in the CRISP-DM Data Preparation and Modeling phases. Many studies also lacked clear frameworks during the business Understanding phase, further weakening the prospects for clinical translation. Inadequate reporting of preprocessing and validation increases the risk of overfitting and reduces reproducibility, while neglect of fairness considerations may perpetuate social biases [19]. To address these issues, future research should adhere to TRIPOD+AI, publish relevant artifacts (e.g., preprocessing pipelines, fairness audits), and systematically address data imbalance to enhance clinical reliability. Standardized reporting is essential to align the rapid growth of the field with higher methodological quality. Our analysis reveals that the field is both maturing and globalizing, but this growth is marked by significant imbalance in its geographic and methodological focus, potentially producing a skewed evidence base [20,21]. The marked surge in publications since 2021, driven by improved data access and the impacts of coronavirus disease 2019, demonstrates rapid expansion [22,23]. However, this growth remains concentrated in certain countries and is heavily focused on supervised learning algorithms. This focus may bias the clinical questions addressed and the solution strategies chosen from a CRISP-DM business understanding perspective, thereby limiting global generalizability and methodological diversity. On a positive note, the strong presence of nurses—with over two-thirds of studies including at least 50% nursing authors—remains a notable strength [24]. Direct nursing involvement is critical for ensuring that research addresses authentic clinical problems and that models are designed with practical workflows in mind, thus improving clinical relevance and the likelihood of successful implementation. The prominence of journals like CIN: Computers, Informatics, Nursing underscores the field’s shift toward digital methods but also highlights the need for broader dissemination across a wider spectrum of clinical and general nursing journals.
In the business understanding phase, our findings indicate that research has overwhelmingly focused on fundamental, high-impact clinical challenges: pressure ulcers, falls, and hospital readmissions. These topics not only appear frequently in the literature but also represent core nursing-sensitive outcomes where ML has a clear potential to enhance patient safety and optimize care quality. For example, the sustained emphasis on pressure ulcer prediction, from early explorations [25] to recent, more sophisticated applications [26], reflects an ongoing effort to shift from reactive treatment to proactive prevention. Similarly, the evolution of fall prediction models from initial systems [27] to tailored applications in nursing homes [28] demonstrates the field’s progression toward targeting high-risk populations. Studies of readmission prediction, spanning patient groups from adults with diabetes [29] to pediatric populations [30], further illustrate the strategic use of ML to advance system-level goals such as cost reduction and continuity of care. Despite this clear clinical focus, a substantial gap exists in the formal use of data mining frameworks. Although CRISP-DM is the most frequently cited methodology, indicating recognition of the need for structured approaches [13], its principles were rarely applied thoroughly. This reveals a frequent disconnect between business understanding and data understanding, undermining the foundations for effective modeling and deployment.
A central paradox emerged in the Data Preparation and Modeling phases: although studies increasingly employ sophisticated techniques, progress was undermined by a persistent lack of reporting transparency, which threatens reproducibility. In the data preparation phase, for instance, researchers used a wide array of methods, ranging from standard encoding to advanced multiple imputation [31,32], and analyzed complex variables for predicting outcomes such as infections and mental health. In the modeling phase, random forests remained the most popular algorithm, recognized for their robust performance with complex data [33], alongside a wide array of other models from interpretable logistic regression to various boosting algorithms. However, this methodological sophistication was not matched by reporting rigor. Our TRIPOD+AI analysis identified transparency gaps that directly compromise reproducibility: for example, only 48.0% of studies reported on missing data handling (Item 11), and just 22.4% described their approach to class imbalance (Item 13), which is a crucial aspect for many clinical outcomes. This deficit extended to the modeling phase, where hyperparameter tuning was often limited to basic grid search [34], and no study reported plans for model recalibration (Item 12f) to address performance drift after deployment. This disconnect between methodological application and transparent reporting severely restricts replication and obscures risks, such as those related to class imbalance in rare but important clinical outcomes. Bringing these methodological threads together, we propose that future nursing ML studies explicitly map every methodological decision to the relevant CRISP-DM phase and publish corresponding artifacts, such as problem definition sheets, EDA dashboards, preprocessing pipelines, tuning logs, and drift-monitoring plans. This level of transparency will enhance reproducibility, facilitate peer auditing, and accelerate clinical translation.
In the evaluation phase, our analysis shows that nursing ML research continues to prioritize algorithmic performance over clinical interpretability, thus limiting translational potential. The reliance on global performance metrics, primarily AUC-ROC and accuracy, demonstrates an emphasis on overall model correctness. This pattern aligns with the frequent identification of ensemble methods like random forest as the top-performing algorithm, praised for their high predictive accuracy on complex nursing datasets [33]. Yet, this focus on aggregate performance can obscure clinical utility. For example, greater attention should be paid to metrics such as sensitivity and specificity, which often hold more clinical significance (e.g., minimizing false negatives for high-risk conditions). This narrow emphasis is compounded by a lack of model interpretability: most studies limited themselves to simple performance comparisons, with relatively little use of tools such as SHAP to explain predictions. This failure to prioritize explainability remains a major barrier to building clinical trust and ensuring that model decisions are safe and equitable. By focusing narrowly on a limited set of performance metrics—without sufficient regard for explainability or fairness audits—current practice falls short of fully satisfying the goals of the CRISP-DM evaluation phase, which should integrate model assessment with broader objectives for patient safety and health equity.
The most significant gap identified in this review lies in the deployment phase, as the vast majority of studies do not progress beyond model evaluation to real-world clinical implementation. Notably, only seven of the 125 analyzed studies reported any form of practical deployment, highlighting a critical research-to-practice gap. This scarcity reflects the immense challenges of clinical integration, which go far beyond model development and require addressing complex safety, reliability, and ethical issues [35]. While hospital-based research remains dominant, recent expansion into home and community care settings is a promising trend. Nevertheless, to bridge the divide between high-performing models and tangible patient impact, future work must prioritize implementation science. This includes developing user-centered evaluation standards and strengthening research on embedding ML tools into diverse clinical workflows to genuinely improve care quality and patient safety.
Concrete solutions include participatory design, real-time missing-data pipelines, and cluster randomized controlled trials. To translate predictive performance into bedside impact, we propose a three-tiered roadmap: (1) Participatory co-design, which involves engaging bedside nurses in early prototype testing to align alert frequency with cognitive load; (2) Integration with EHR and Clinical Decision Support, by leveraging Fast Healthcare Interoperability Resources-based Application Programming Interfaces so that model outputs populate existing decision-support widgets instead of separate dashboards; and (3) Prospective hybrid trials, combining A/B-tested usability endpoints with cluster randomized outcome metrics to evaluate both adoption and effectiveness. Key barriers include data governance concerns, alert fatigue, and algorithmic bias. Mitigation strategies may include federated learning to address data privacy, threshold-adaptive alerting, and continuous fairness audits.
In summary, when viewed through the CRISP-DM framework, methodological weaknesses accumulate across phases—from insufficient problem formalization and superficial data exploration, to opaque preprocessing, limited hyperparameter optimization, narrow evaluation, and minimal deployment planning. These limitations constrain the real-world impact of nursing ML. Addressing them will require transparent protocols, fairness-aware analytics, robust tuning and monitoring, and rigorous clinical trials to ensure that predictive models ultimately translate into improved patient outcomes and nursing practice.
Despite the review’s methodological rigor, several limitations should be acknowledged. First, the definition of 'nursing research' was operationalized by requiring at least one author with a nursing affiliation. While this provided an objective and reproducible screening criterion, it has limitations as a proxy for direct relevance to nursing practice. This approach may have excluded valuable interdisciplinary studies where ML was applied to nursing-sensitive outcomes (e.g., patient falls, pressure injuries) but conducted by teams lacking a formally affiliated nursing researcher. Conversely, it may have included studies where a nursing author's involvement was minimal and the research focus was not central to clinical nursing. Future reviews could use a more nuanced, two-stage approach: an initial broad search for nursing-sensitive outcomes, followed by a content-based assessment of each study’s direct applicability to nursing practice, though this would introduce greater subjectivity into the selection process. Second, while the review team’s nursing background ensured clinical relevance, their interpretations of ML methods and performance may reflect a nursing-centered perspective; researchers from computer science or ML fields might have offered different evaluations of model selection, hyperparameter tuning, or performance metrics. Third, although the TRIPOD+AI checklist is suitable for evaluating clinical prediction models, it may not fully capture the methodological nuances of nursing research, particularly in studies involving exploratory or unsupervised approaches. The moderate mean compliance rate of 50.4% underscores broader issues in methodological rigor and reporting transparency. Fourth, our deliberate exclusion of DL studies, while necessary for methodological consistency, means that this review does not represent the entire landscape of AI in nursing. The rapidly growing body of research utilizing CNNs for medical imaging or RNNs for sequential EHR analysis falls outside the scope of this review. Consequently, our findings and conclusions are specific to traditional ML applications, and a separate, dedicated systematic review is needed to characterize the unique methodologies and challenges of DL in the nursing field.
Nevertheless, this review has notable strengths. It is the first synthesis to apply the TRIPOD+AI checklist and PRISMA framework to 125 nursing ML studies, anchored in a pre-registered PROSPERO protocol and structured by the CRISP-DM model. The search strategy encompassed nine international and five Korean databases, thereby minimizing language and regional bias. Dual independent screening and quality appraisal further reduced reviewer subjectivity, while the large sample size enabled robust identification of reporting and algorithmic trends spanning two decades. By linking checklist findings to phase-specific recommendations, this review provides actionable guidance for future nursing ML research and editorial policy development.
This study systematically reviewed and analyzed the application of ML in nursing, highlighting both its achievements and limitations. The use of ML in nursing research has recently increased rapidly, with a predominant focus on patient safety and healthcare quality improvement—particularly in areas such as pressure ulcer prediction, fall prevention, and hospital readmission management. Algorithms such as random forest, XGBoost, and logistic regression were widely employed, with performance metrics like AUC-ROC and accuracy most commonly used for evaluation. However, several factors limit comparability and reproducibility across studies, including inadequate reporting of data preprocessing methods, inconsistent performance evaluation criteria, and insufficient attention to algorithm fairness and ethical considerations. The generally low compliance rate with the TRIPOD+AI checklist underscores the need to improve transparency and reliability in nursing ML research. Additional shortcomings included insufficient detail on hyperparameter tuning and model performance evaluation processes, as well as limited real-world deployment of ML tools. Consequently, there is an urgent need to standardize research design and reporting practices to enhance the quality of ML studies in nursing. Adherence to structured reporting guidelines, such as TRIPOD+AI, can significantly improve transparency and reproducibility. Furthermore, future research should prioritize practical deployment in diverse clinical settings, ongoing model performance optimization, and fairness assurance to strengthen clinical applicability. By adopting systematic and standardized approaches, future nursing ML research can enhance practical relevance and contribute to improved patient-centered nursing care quality.

CONFLICTS OF INTEREST

The authors declared no conflict of interest.

AUTHORSHIP

Study conception and/or design acquisition - EJK and SKK; analysis - EJK and SKK; interpretation of the data - EJK and SKK; and drafting or critical revision of the manuscript for important intellectual content - EJK and SKK.

FUNDING

None.

ACKNOWLEDGEMENT

None.

DATA AVAILABILITY STATEMENT

No new data were created or analyzed during this study. Data sharing is not applicable to this article.

Supplementary materials can be found via https://doi.org/10.7475/kjan.2025.0327.

Supplement Table 1.

Presenting the PICOTS-SD Process Applied in the Screening Step
kjan-2025-0327-Supplemental-Table-1.pdf

Supplement Table 2.

Analysis of Business Understanding phase based on CRISP-DM methodology in selected studies (N=125)
kjan-2025-0327-Supplemental-Table-2.pdf

Supplement Table 3.

Results Based on Data Understanding in CRISP-DM (N=125)
kjan-2025-0327-Supplemental-Table-3.pdf

Supplement Table 4.

Results Based on Data Preparation in CRISP-DM (N=125)
kjan-2025-0327-Supplemental-Table-4.pdf

Supplement Table 5.

Results Based on Modeling in CRISP-DM (N=125)
kjan-2025-0327-Supplemental-Table-5.pdf

Supplement Table 6.

Results Based on Evaluation in CRISP-DM (N=125)
kjan-2025-0327-Supplemental-Table-6.pdf

Supplement Table 7.

Results based on Deployment in CRISP-DM (N=125)
kjan-2025-0327-Supplemental-Table-7.pdf
Figure 1.
PRISMA flow diagram.
kjan-2025-0327f1.jpg
Table 1.
CRISP-DM Process Model Descriptions
No. Phases Short descriptions
1 Business understanding - Determine business objectives, assess situation, define data mining goals, develop project plan.
Understanding business goals and translating them into data mining objectives.
2 Data understanding - Initial data collection, data description, data exploration, data quality assessment.
Collecting data, familiarizing oneself with data, identifying quality issues, and gaining initial insights.
3 Data preparation - Data selection, data cleaning, data construction, data integration, data formatting.
All activities required to construct the final dataset from initial raw data.
4 Modeling - Select modeling techniques, design tests, build models, evaluate models.
Selecting, applying, and optimizing modeling techniques.
5 Evaluation - Evaluate results, review processes, determine next steps.
Evaluating the model from the perspective of achieving business objectives and reviewing the entire process.
6 Deployment - Plan deployment, monitor and maintain the model, produce final reports, review project.
Integrating the model into actual business processes and ensuring organizational usage of the outcomes.

CRISP-DM=Cross-Industry Standard Process for Data Mining.

Table 2.
Data Extraction Plan
No. Phases Data extraction details
1 General characteristics 1) Journal name, 2) Year of publication, 3) Authors, 4) Proportion of nursing-affiliated authors (determined based on institutional affiliation), 5) Country, 6) Type of ML (supervised/unsupervised/reinforcement), 7) Type of algorithm used (prediction/classification/clustering, etc.)
2 Business understanding 8) Research objective, 9) Research design and methodology (KDD, CRISP-DM, etc.)
3 Data understanding 10) Tools and software used for data analysis, 11) Data source, 12) EDA methods, 13) Target of ML application, 14) Sample size
4 Data preparation 15) Data preprocessing techniques (normalization, standardization, encoding, etc.), 16) Predictor (explanatory/training) variables, 17) Number of predictor variables, 18) Target variable (the variable to be predicted, classified, or analyzed), 19) Data split ratio
5 Modeling 20) Applied ML algorithms and modeling techniques, 21) Hyperparameter tuning methods
6 Evaluation 22) Confusion matrix, 23) Performance evaluation metrics for classification and regression models, 24) Performance evaluation results for each model, 25) Best model, 26) Method of analyzing variable importance or the impact on the target variable (feature importance, SHAP value, etc.)
7 Deployment 27) Research environment, 28) IRB approval, 29) Whether the model was actually deployed

CRISP-DM=Cross-Industry Standard Process for Data Mining; EDA=exploratory data analysis; IRB=Institutional Review Board; KDD=Knowledge Discovery in Databases; ML=machine learning; SHAP=Shapley additive explanations.

Table 3.
General Characteristics of the Selected Studies (N=125)
Characters Categories n (%)
Location† Asia 71 (35.9)
North America 59 (29.8)
Europe 22 (11.1)
Oceania 4 (2.0)
South America 2 (1.0)
Publication year Before 2020 26 (20.8)
2020 11 (8.8)
2021 20 (16.0)
2022 18 (14.4)
2023 24 (19.2)
2024 26 (20.8)
Percentage of nurses on the research team (%) 0.0–0.25 16 (12.8)
0.25–0.5 26 (20.8)
0.5–0.75 41 (32.8)
0.75–1.0 42 (33.6)
ML type Supervised learning 117 (93.6)
Unsupervised learning 5 (4.0)
Supervised and unsupervised learning 3 (2.4)
Algorithm type Classification 101 (80.8)
Regression 14 (11.2)
Clustering 7 (5.6)
Classification, association rule mining 2 (1.6)
Regression, dimensionality reduction 1 (0.8)
Research objective Pressure-injury/ulcer 24 (19.2)
Readmission/utilization 17 (13.6)
Fall risk 10 (8.0)
Others 74 (59.2)
Journal CIN: Computers, Informatics, Nursing 7 (5.6)
Applied Clinical Informatics 4 (3.2)
International Journal of Environmental Research and Public Health 4 (3.2)
Journal of the American Medical Informatics Association 4 (3.2)
BMC Medical Informatics and Decision Making 3 (2.4)
International Journal of Medical Informatics 3 (2.4)
Journal of Advanced Nursing 3 (2.4)
Journal of Nursing Management 3 (2.4)
Nursing Research 3 (2.4)
BMC Nursing 2 (1.6)
Journal of Biomedical Informatics 2 (1.6)
Journal of Emergency Nursing 2 (1.6)
Journal of Medical Internet Research 2 (1.6)
Journal of Nursing Scholarship 2 (1.6)
Journal of the American Medical Directors Association 2 (1.6)
International Journal of Nursing Studies 2 (1.6)
Healthcare 2 (1.6)
Innovation in Applied Nursing Informatics 2 (1.6)
Nurse Education Today 2 (1.6)
Archives of Psychiatric Nursing 2 (1.6)
Others 69 (55.2)

ML=machine learning;

†The number of “location” values exceeds the 125 included studies because location data were compiled for each author, and a single study could involve authors from multiple locations.

Table 4.
Summary of the Findings of This Study
Sections Key findings
Quality appraisal of the studies The TRIPOD+AI appraisal showed moderate compliance (≈50%). Core methods were well reported (>85%), but transparency and ethical aspects were weak (<25%).
General characteristics of the selected studies USA (30.8%), China (14.5%), and Korea (10.7%) dominated. Most studies used supervised learning (93.6%), especially classification tasks (80.8%).
Problem definition for research objective Top topics were pressure injury, fall, and readmission. Most studies lacked formal data-mining frameworks; CRISP-DM was the most used among those that did.
Data collection and exploration R, Python, and SPSS were the most used tools. EMR/EHR and survey data dominated. Sample sizes ranged from 5 to over 3.5 million cases.
Data preparation Common preprocessing included standardization, normalization, and imputation. Label encoding and one-hot encoding were frequent. Most studies used <50 predictors.
Model building Random forest was the most used algorithm, followed by logistic regression, SVM, and XGBoost. Hyperparameter tuning was often omitted; Grid Search was most common when used.
Evaluation and review AUC-ROC, accuracy, sensitivity, and F1-score were most reported. RF was most often top-performing; 42 studies did not report variable importance.
Deployment Hospitals were the most common setting. IRB approval was reported in 76.0% of studies. Only seven studies described actual deployment.

AUC-ROC=area under the receiver operating characteristic curve; CRISP-DM=Cross-Industry Standard Process for Data Mining; EHR=electronic health record; EMR=electronic medical record; IRB=Institutional Review Board; RF=random forest; SVM=support vector machine; TRIPOD+AI=Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis+Artificial Intelligence extension; XGBoost=eXtreme gradient boosting.

Appendix 1.
List of studies included in the systematic review
A1. Hiissa M, Marketta, et al. Towards automated classification of intensive care nursing narratives. In: Ubiquity: Technologies for Better Health in Aging Societies. Amsterdam (Netherlands): IOS Press; 2006. p. 789–794.
A2. Nii M, et al. Nursing-care freestyle text classification using support vector machines. In: 2007 IEEE International Conference on Granular Computing (GRC 2007). Piscataway (NJ): IEEE; 2007. doi: 10.1109/GrC.2007.131
A3. Moseley LG, Mead DM. Predicting who will drop out of nursing courses: a machine learning exercise. Nurse Educ Today. 2008;28(4):469–75. doi: 10.1016/j.nedt.2007.07.012
A4. Zlotnik A, et al. Emergency department visit forecasting and dynamic nursing staff allocation using machine learning techniques with readily available open-source software. CIN: Comput Inform Nurs. 2015;33(8):368–77. doi: 10.1097/CIN.0000000000000173
A5. Nii M, et al. Nursing-care text evaluation using word vector representations realized by word2vec. In: 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE). Piscataway (NJ): IEEE; 2016. doi: 10.1109/FUZZ-IEEE.2016.7737960
A6. Sherwin R, Ying H, Kakarla P. 31 performance of a novel computer-based clinical decision support alert and the impact of patient partitioning and optimization to identify septic patients in an urban emergency department. Ann Emerg Med. 2017;70(4 Suppl):S13.
A7. Yokota S, Endo M, Ohe K. Establishing a classification system for high fall-risk among inpatients using support vector machines. CIN: Comput Inform Nurs. 2017;35(8):408–16. doi: 10.1097/CIN.0000000000000332
A8. Choi J, Choi J, Choi WJ. Predicting depression among community residing older adults: a use of machine learning approach. In: Nursing Informatics 2018. Amsterdam (Netherlands): IOS Press; 2018. p. 265.
A9. Culliton P, et al. Predicting severe sepsis using text from the electronic health record [Preprint]. arXiv:1711.11536 [cs.CL]. 2017. Available from: https://doi.org/10.48550/arXiv.1711.11536
A10. Bose E, et al. Machine learning methods for identifying critical data elements in nursing documentation. Nurs Res. 2019;68(1):65–72. doi: 10.1097/NNR.0000000000000315
A11. Gannod GC, et al. A machine learning recommender system to tailor preference assessments to enhance person-centered care among nursing home residents. Gerontologist. 2019;59(1):167–76. doi: 10.1093/geront/gny056
A12. Johnson SG, Pruinelli L, Westra BL. Machine learned mapping of local EHR flowsheet data to standard information models using topic model filtering. AMIA Annu Symp Proc. 2020;2019:504-513.
A13. Korach ZT, et al. Unsupervised machine learning of topics documented by nurses about hospitalized patients prior to a rapid-response event. Appl Clin Inform. 2019;10(5):952–63. doi: 10.1055/s-0039-3401814
A14. Kwon JY, et al. Nurses ‘seeing forest for the trees’ in the age of machine learning: using nursing knowledge to improve relevance and performance. CIN: Comput Inform Nurs. 2019;37(4):203–12. doi: 10.1097/CIN.0000000000000508
A15. Sullivan SS, et al. Mortality risk in homebound older adults predicted from routinely collected nursing data. Nurs Res. 2019;68(2):156–66. doi: 10.1097/NNR.0000000000000328
A16. Topaz M, et al. Mining fall-related information in clinical notes: comparison of rule-based and novel word embedding-based machine learning approaches. J Biomed Inform. 2019;90:103103. doi: 10.1016/j.jbi.2019.103103
A17. Brom H, et al. Leveraging electronic health records and machine learning to tailor nursing care for patients at high risk for readmissions. J Nurs Care Qual. 2020;35(1):27–33. doi: 10.1097/NCQ.0000000000000412
A18. Fritz RL, et al. Automated smart home assessment to support pain management: multiple methods analysis. J Med Internet Res. 2020;22(11):e23943. doi: 10.2196/23943
A19. Horvat CM, et al. Development and initial implementation of a machine-learning-based predictive index for critical deterioration among hospitalized children. Pediatrics. 2020;146(1_MeetingAbstract):11–2. doi: 10.1542/peds.146.1MA1.11
A20. Hu R, et al. Using machine learning to identify top predictors for nurses’ willingness to report medication errors. Array. 2020;8:100049. doi: 10.1016/j.array.2020.100049
A21. Ladios-Martin M, et al. Predictive modeling of pressure injury risk in patients admitted to an intensive care unit. Am J Crit Care. 2020;29(4):e70–80. doi: 10.4037/ajcc2020237
A22. Lee SK, et al. Application of machine learning methods in nursing home research. Int J Environ Res Public Health. 2020;17(17):6234. doi: 10.3390/ijerph17176234
A23. Liang C, et al. Toward systems-centered analysis of patient safety events: improving root cause analysis by optimized incident classification and information presentation. Int J Med Inform. 2020;135:104054. doi: 10.1016/j.ijmedinf.2019.104054
A24. Lindberg DS, et al. Identification of important factors in an inpatient fall risk prediction model to improve the quality of care using EHR and electronic administrative data: a machine-learning approach. Int J Med Inform. 2020;143:104272. doi: 10.1016/j.ijmedinf.2020.104272
A25. Park JI, et al. Knowledge discovery with machine learning for hospital-acquired catheter-associated urinary tract infections. CIN: Comput Inform Nurs. 2020;38(1):28–35. doi: 10.1097/CIN.0000000000000562
A26. Topaz M, et al. Home healthcare clinical notes predict patient hospitalization and emergency department visits. Nurs Res. 2020;69(6):448–54. doi: 10.1097/NNR.0000000000000470
A27. Womack DM, et al. Registered nurse strain detection using ambient data: an exploratory study of underutilized operational data streams in the hospital workplace. Appl Clin Inform. 2020;11(4):598–605. doi: 10.1055/s-0040-1715829
A28. An R, et al. Machine learning-based patient classification system for adult patients in intensive care units: a cross-sectional study. J Nurs Manag. 2021;29(6):1752–62. doi: 10.1111/jonm.13284
A29. Chen L. Facial expression recognition with machine learning and assessment of distress in patients with cancer. Oncol Nurs Forum. 2021;48(1):81–93. doi: 10.1188/21.ONF.81-93
A30. Conway A, et al. Predicting prolonged apnea during nurse-administered procedural sedation: machine learning study. JMIR Perioper Med. 2021;4(2):e29200. doi: 10.2196/29200
A31. Garcés-Jiménez A, et al. Medical prognosis of infectious diseases in nursing homes by applying machine learning on clinical data collected in cloud microservices. Int J Environ Res Public Health. 2021;18(24):13278. doi: 10.3390/ijerph182413278
A32. Hannaford L, Cheng X, Kunes-Connell M. Predicting nursing baccalaureate program graduates using machine learning models: a quantitative research study. Nurse Educ Today. 2021;99:104784. doi: 10.1016/j.nedt.2021.104784
A33. Havaei F, et al. Identifying the most important workplace factors in predicting nurse mental health using machine learning techniques. BMC Nurs. 2021;20:1–10. doi: 10.1186/s12912-021-00742-9
A34. Howard EP, et al. Machine-learning modeling to predict hospital readmission following discharge to post-acute care. J Am Med Dir Assoc. 2021;22(5):1067–72. doi: 10.1016/j.jamda.2020.12.017
A35. Hu M, et al. A risk prediction model based on machine learning for cognitive impairment among Chinese community-dwelling elderly people with normal cognition: development and validation study. J Med Internet Res. 2021;23(2):e20298. doi: 10.2196/20298
A36. Ivanov O, et al. Improving ED emergency severity index acuity assignment using machine learning and clinical natural language processing. J Emerg Nurs. 2021;47(2):265–78. doi: 10.1016/j.jen.2020.11.001
A37. Jin L, et al. Intervention prediction for patients with pressure injury using random forest. In: 2021 IEEE International Conference on Big Knowledge (ICBK). Piscataway (NJ): IEEE; 2021. doi: 10.1109/ICKG52313.2021.00072
A38. Kim J, Jang I. Predictors of bleeding event among elderly patients with mechanical valve replacement using random forest model: a retrospective study. Medicine (Baltimore). 2021;100(19):e25875. doi: 10.1097/MD.0000000000025875
A39. Lee SK, et al. Identifying the risk factors associated with nursing home residents’ pressure ulcers using machine learning methods. Int J Environ Res Public Health. 2021;18(6):2954. doi: 10.3390/ijerph18062954
A40. Liu CH, Hu YH, Lin YH. A machine learning-based fall risk assessment model for inpatients. CIN: Comput Inform Nurs. 2021;39(8):450–9. doi: 10.1097/CIN.0000000000000727
A41. Macieira TGR, Yao Y, Keenan GM. Use of machine learning to transform complex standardized nursing care plan data into meaningful research variables: a palliative care exemplar. J Am Med Inform Assoc. 2021;28(12):2695–701. doi: 10.1093/jamia/ocab205
A42. Nagata T, et al. Skin tear classification using machine learning from digital RGB image. J Tissue Viability. 2021;30(4):588–93. doi: 10.1016/j.jtv.2021.01.004
A43. Nakagami G, et al. Supervised machine learning-based prediction for in-hospital pressure injury development using electronic health records: a retrospective observational cohort study in a university hospital in Japan. Int J Nurs Stud. 2021;119:103932. doi: 10.1016/j.ijnurstu.2021.103932
A44. Song W, et al. Predicting pressure injury using nursing assessment phenotypes and machine learning methods. J Am Med Inform Assoc. 2021;28(4):759–65. doi: 10.1093/jamia/ocaa336
A45. Yang R, et al. Predicting falls among community-dwelling older adults: a demonstration of applied machine learning. CIN: Comput Inform Nurs. 2021;39(5):273–80. doi: 10.1097/CIN.0000000000000688
A46. Zhou H, et al. Using machine learning to predict paediatric 30-day unplanned hospital readmissions: a case-control retrospective analysis of medical records, including written discharge documentation. Aust Health Rev. 2021;45(3):328–37.
A47. Dong YH, et al. Investigating psychological differences between nurses and other health care workers from the Asia-Pacific region during the early phase of COVID-19: machine learning approach. JMIR Nurs. 2022;5(1):e32647. doi: 10.2196/32647
A48. Havaei F, Ji XR, Boamah SA. Workplace predictors of quality and safe patient care delivery among nurses using machine learning techniques. J Nurs Care Qual. 2022;37(2):103–9. doi: 10.1097/NCQ.0000000000000600
A49. Hu T, et al. Establishment of a model for predicting the outcome of induced labor in full-term pregnancy based on machine learning algorithm. Sci Rep. 2022;12(1):19063. doi: 10.1038/s41598-022-21954-2
A50. Jin S, et al. Machine learning predicts cancer-associated deep vein thrombosis using clinically available variables. Int J Med Inform. 2022;161:104733. doi: 10.1016/j.ijmedinf.2022.104733
A51. Ladios-Martin M, Cabañero-Martínez MJ, Fernández-de-Maya J, et al. Development of a predictive inpatient falls risk model using machine learning. J Nurs Manag. 2022;30(8):3777-3786. doi: 10.1111/jonm.13760.
A52. Lee YJ, et al. Identifying language features associated with needs of ovarian cancer patients and caregivers using social media. Cancer Nurs. 2022;45(3):E639–45. doi: 10.1097/NCC.0000000000000928
A53. Mishra AK, et al. Explainable fall risk prediction in older adults using gait and geriatric assessments. Front Digit Health. 2022;4:869812. doi: 10.3389/fdgth.2022.869812
A54. Moon KJ, et al. The development of a web-based app employing machine learning for delirium prevention in long-term care facilities in South Korea. BMC Med Inform Decis Mak. 2022;22(1):220. doi: 10.1186/s12911-022-01966-8
A55. Padhye N, et al. Pressure injury link to entropy of abdominal temperature. Entropy (Basel). 2022;24(8):1127. doi: 10.3390/e24081127
A56. Qian D, Gao H. Efficacy analysis of team-based nursing compliance in young and middle-aged diabetes mellitus patients based on random forest algorithm and logistic regression. Comput Math Methods Med. 2022;2022:3882425. doi: 10.1155/2022/3882425
A57. Rojo J, et al. Improving the assessment of older adults using feature selection and machine learning models. Gerontechnology. 2022;21(s):1-1. doi: 10.4017/gt.2022.21.s.544.opp4
A58. Song J, et al. Clinical notes: an untapped opportunity for improving risk prediction for hospitalization and emergency department visit during home health care. J Biomed Inform. 2022;128:104039. doi: 10.1016/j.jbi.2022.104039
A59. Spiller TR, et al. Delirium screening in an acute care setting with a machine learning classifier based on routinely collected nursing data: a model development study. J Psychiatr Res. 2022;156:194–9. doi: 10.1016/j.jpsychires.2022.10.018
A60. Walker K, et al. Emergency medicine patient wait time multivariable prediction models: a multicentre derivation and validation study. Emerg Med J. 2022;39(5):386–93. doi: 10.1101/2021.03.19.21253921
A61. Widyawati MN, Astuti EHP. Human-in-the-loop application design for early detection of pregnancy danger signs. Belitung Nurs J. 2022;8(2):161. doi: 10.33546/bnj.1984
A62. Xu J, et al. Development and validation of a machine learning algorithm–based risk prediction model of pressure injury in the intensive care unit. Int Wound J. 2022;19(7):1637–49. doi: 10.1111/iwj.13764
A63. Yakusheva O, et al. Nonlinear association of nurse staffing and readmissions uncovered in machine learning analysis. Health Serv Res. 2022;57(2):311–21. doi: 10.1111/1475-6773.13695
A64. Aydın AI, Özyazıcıoğlu N. Assessment of postoperative pain in children with computer assisted facial expression analysis. J Pediatr Nurs. 2023;71:60–5. doi: 10.1016/j.pedn.2023.03.008
A65. Rui C, et al. Construction and validation of a model for predicting the risk of immune checkpoint inhibitor pneumonitis. Chin J Pract Nurs. 2023;36:2458–64.
A66. Chen YH, Xu JL. Applying artificial intelligence to predict falls for inpatient. Front Med (Lausanne). 2023;10:1285192. doi: 10.3389/fmed.2023.1285192
A67. Dai PY, et al. Agitation sedation monitoring system for intensive care unit based on ensemble learning model. In: Proceedings of the 2023 7th International Conference on Medical and Health Informatics. New York (NY): Association for Computing Machinery; 2023. doi: 10.1145/3608298.3608334
A68. Dweekat OY, Lam SS, McGrath L. A hybrid system of Braden scale and machine learning to predict hospital-acquired pressure injuries (bedsores): a retrospective observational cohort study. Diagnostics (Basel). 2022;13(1):31. doi: 10.3390/diagnostics13010031
A69. Edgcomb J, et al. Computable phenotyping of children with suicide-related emergencies using nursing triage safety screening and interventions. In: AACAP’s 70th Annual Meeting; 2023. [conference abstract].
A70. Gajra A, et al. Reducing avoidable emergency visits and hospitalizations with patient risk-based prescriptive analytics: a quality improvement project at an oncology care model practice. JCO Oncol Pract. 2023;19(5):e725–31. doi: 10.1200/OP.22.00307
A71. Havaei F, et al. Workplace predictors of violence against nurses using machine learning techniques: a cross-sectional study utilizing the national standard of psychological workplace health and safety. Healthcare (Basel). 2023;11(7):1008. doi: 10.3390/healthcare11071008
A72. Hewner S, Smith E, Sullivan SS. Identifying high-need primary care patients using nursing knowledge and machine learning methods. Appl Clin Inform. 2023;14(3):408–17. doi: 10.1055/a-2048-7343
A73. Im S, Lee S. Development of mortality prediction model using electronic health record (EHR) data and machine learning algorithm in intensive care unit (ICU). J Korean Data Anal Soc. 2023 Oct;25(5):1977–92. doi: 10.37727/jkdas.2023.25.5.1977
A74. Kim J, et al. Identifying the suicidal ideation risk group among older adults in rural areas: developing a predictive model using machine learning methods. J Adv Nurs. 2023;79(2):641–51. doi: 10.1111/jan.15549
A75. Kim SK, et al. Development of a nurse turnover prediction model in Korea using machine learning. Healthcare (Basel). 2023;11(11):1583. doi: 10.3390/healthcare11111583
A76. Lee H, Lee S, Kim H. Factors affecting the length of stay in the emergency department for critically ill patients transferred to regional emergency medical center. Nurs Open. 2023;10(5):3220–31. doi: 10.1002/nop2.1573
A77. Lee H, et al. Predicting workplace violence in the emergency department based on electronic health record data. J Emerg Nurs. 2023;49(3):415–24. doi: 10.1016/j.jen.2023.01.010
A78. Lee LL, Chen SL. The application of hyperspectral imaging to the measurement of pressure injury area. Int J Environ Res Public Health. 2023;20(4):2851. doi: 10.3390/ijerph20042851
A79. Li L, et al. Severe hypernatremia during postoperative care in patients with craniopharyngioma. Endocr Connect. 2023;12(12):e230149. doi: 10.1530/EC-23-0149
A80. Liao PH, et al. Application of machine learning and its effects on the development of a nursing guidance mobile app for sarcopenia. BMC Nurs. 2023;22(1):369. doi: 10.1186/s12912-023-01545-w
A81. Martha SR, et al. Machine learning analysis of the cerebrovascular thrombi lipidome in acute ischemic stroke. J Neurosci Nurs. 2023;55(1):10–7. doi: 10.1097/JNN.0000000000000682
A82. Shanmugam AJ, Hamid HIA, Dailah HGH, Begum SK, Ahamed SB. Analysis of undergraduate student’s knowledge of self-medication practice using machine learning algorithms. SSRG Int J Electr Electron Eng. 2023 May;10(5):60–8. doi: 10.14445/23488379/IJEEE-V10I5P106
A83. Rodríguez-Vico A, et al. Predictores del estado post-ictus en el alta hospitalaria. Importancia en enfermería. Enferm Glob. 2023;22(69):1–37. doi: 10.6018/eglobal.530591
A84. Yan Z, et al. Construction and validation of machine learning algorithms to predict chronic post-surgical pain among patients undergoing total knee arthroplasty. Pain Manag Nurs. 2023;24(6):627–33. doi: 10.1016/j.pmn.2023.04.008
A85. Yıldız M, et al. The effect of intercultural sensitivity and ethnocentrism on health tourism awareness level in nurses: analysis with machine learning approach. Arch Psychiatr Nurs. 2023;46:40–50. doi: 10.1016/j.apnu.2023.07.002
A86. Zhou Y, et al. Developing a machine learning model for detecting depression, anxiety, and apathy in older adults with mild cognitive impairment using speech and facial expressions: a cross-sectional observational study. Int J Nurs Stud. 2023;146:104562. doi: 10.1016/j.ijnurstu.2023.104562
A87. Zolnoori M, et al. Is the patient speaking or the nurse? Automatic speaker type identification in patient–nurse audio recordings. J Am Med Inform Assoc. 2023;30(10):1673–83. doi: 10.1093/jamia/ocad139
A88. Park YT, Lee SM, Lee YH, Kim KG. Performance evaluation of artificial intelligence methods predicting annual number of patients in hospitals. HIRA Res. 2024;4(1):73–86. doi: 10.52937/hira.24.4.1.e4
A89. Abi Khalil C, et al. Evaluation of machine learning algorithms for pressure injury risk assessment in a hospital with limited IT resources. In: Digital Health and Informatics Innovations for Sustainable Health Care Systems. Amsterdam (Netherlands): IOS Press; 2024. p. 1033–7. doi: 10.3233/SHTI240587
A90. Alqarrain Y, et al. Data preparation for supervised learning: improving nursing situation awareness to reduce healthcare-acquired urinary tract infection. In: Innovation in Applied Nursing Informatics. Amsterdam (Netherlands): IOS Press; 2024. p. 305–10. doi: 10.3233/SHTI240158
A91. Aryal K, et al. Evaluating methods for risk prediction of COVID-19 mortality in nursing home residents before and after vaccine availability: a retrospective cohort study. BMC Med Res Methodol. 2024;24(1):77. doi: 10.1186/s12874-024-02189-3
A92. Chavan R, Dumbre D, Devi S. Predictive modeling for B. Sc. nursing placement using machine learning algorithms. In: 2024 International Conference on Advances in Computing, Communication and Applied Informatics (ACCAI). Piscataway (NJ): IEEE; 2024. doi: 10.1109/ACCAI61061.2024.10601953
A93. Chen X, et al. A predictive model of pressure injury in children undergoing living donor liver transplantation based on machine learning algorithm. J Adv Nurs. 2024. Epub ahead of print. doi: 10.1111/jan.16449
A94. Crowe C, et al. Treatment effect analysis of the Frailty Care Bundle (FCB) in a cohort of patients in acute care settings. Aging Clin Exp Res. 2024;36(1):187. doi: 10.1007/s40520-024-02840-5
A95. Dai T, et al. A risk prediction model based on machine learning algorithm for parastomal hernia after permanent colostomy. BMC Med Inform Decis Mak. 2024;24(1):224. doi: 10.1186/s12911-024-02627-8
A96. Duarte M, et al. Prediction of positive Patient Health Questionnaire-2 screening using area deprivation index in primary care. Clin Nurs Res. 2024;33(5):355–69. doi: 10.1177/10547738241252887
A97. Guo YF, et al. Effects of job crafting and leisure crafting on nurses’ burnout: a machine learning-based prediction analysis. J Nurs Manag. 2024;2024:9428519. doi: 10.1155/2024/9428519
A98. Jin R, et al. Fairness in classifying and grouping health equity information. In: Innovation in Applied Nursing Informatics. Amsterdam (Netherlands): IOS Press; 2024. p. 368–72. doi: 10.3233/SHTI240171
A99. Kawashima A, et al. Predictive models for palliative care needs of advanced cancer patients receiving chemotherapy. J Pain Symptom Manage. 2024;67(4):306–16. doi: 10.1016/j.jpainsymman.2024.01.009
A100. Kim Y, Kim Y, Choi M. Machine learning-based prediction models of mortality for intensive care unit patients using nursing records. In: Innovation in Applied Nursing Informatics. Amsterdam (Netherlands): IOS Press; 2024. p. 604–5. doi: 10.3233/SHTI240237
A101. Lee JH, et al. Development of a pressure injury machine learning prediction model and integration into clinical practice: a prediction model development and validation study. Korean J Adult Nurs. 2024;36(3):191–202. doi: 10.7475/kjan.2024.36.3.191
A102. Lee PC, et al. Applying machine learning to construct an association model for lung cancer and environmental hormone high-risk factors and nursing assessment reconstruction. J Nurs Scholarsh. 2024;57(1):140-151. doi: 10.1111/jnu.12997
A103. Manworren RC, et al. Performance evaluation of a supervised machine learning pain classification model developed by neonatal nurses. Adv Neonatal Care. 2024;24(3):301–10. doi: 10.1097/ANC.0000000000001145
A104. Rosa NG, Vaz TA, Lucena AF. Nursing workload: use of artificial intelligence to develop a classifier model. Rev Lat Am Enfermagem. 2024;32:e4239. doi: 10.1590/1518-8345.7131.4239
A105. Scroggins JK, et al. Does synthetic data augmentation improve the performances of machine learning classifiers for identifying health problems in patient–nurse verbal communications in home healthcare settings? J Nurs Scholarsh. 2024;57(1):47–58. doi: 10.1111/jnu.13004
A106. Shao L, Wang Z, Xie X, et al. Development and external validation of a machine learning-based fall prediction model for nursing home residents: a prospective cohort study. J Am Med Dir Assoc. 2024;25:105169. doi: 10.1016/j.jamda.2024.105169
A107. Stanik M, Hass Z, Kong N. Seizure prediction in stroke survivors who experienced an infection at skilled nursing facilities—a machine learning approach. Front Physiol. 2024;15:1399374. doi: 10.3389/fphys.2024.1399374
A108. Tahyudin I, et al. Optimizing stroke mortality prediction: a comprehensive study on risk factors analysis and hyperparameter tuning techniques. TEM J. 2024;13(1):705. doi: 10.18421/TEM131-74
A109. Yıldız M, et al. Investigation the relationship between xenophobic attitude and intercultural sensitivity level in nurses. Arch Psychiatr Nurs. 2024;48:20–9. doi: 10.1016/j.apnu.2023.12.002
A110. Yu C, et al. Stress begets stress: the moderating role of childhood adversity in the relationship between job stress and sleep quality among nurses. J Affect Disord. 2024;348:345–52. doi: 10.1016/j.jad.2023.12.090
A111. Zhang W, et al. Development and validation of machine learning models to predict frailty risk for elderly. J Adv Nurs. 2024;80(12):5064–75. doi: 10.1111/jan.16192
A112. Zolnoori M, et al. Utilizing patient-nurse verbal communication in building risk identification models: the missing critical data stream in home healthcare. J Am Med Inform Assoc. 2024;31(2):435–44. doi: 10.1093/jamia/ocad195
A113. Zhang Y, et al. Nursing factors associated with length of stay and readmission rate of the elderly residents from nursing home based on LTCfocus database. Public Health. 2022;213:19–27. doi: 10.1016/j.puhe.2022.09.011
A114. Erfani G, et al. Identifying patterns and profiles of vaccination hesitancy among nurses for tailoring healthcare policies in the UK: a cross-sectional study. Int Nurs Rev. 2024. doi: 10.1111/inr.13035
A115. Kim TY, Lang N. Predictive modeling for the prevention of hospital-acquired pressure ulcers. AMIA Annu Symp Proc. 2006;2006: 434-438.
A116. Cho IS, Chung E. Predictive Bayesian network model using electronic patient records for prevention of hospital-acquired pressure ulcers. J Korean Acad Nurs. 2011 Jun;41(3):423–31. doi: 10.4040/jkan.2011.41.3.423
A117. Setoguchi Y, Ghaibeh AA, Mitani K, et al. Predictability of pressure ulcers based on operation duration, transfer activity, and body mass index through the use of an alternating decision tree. J Med Invest. 2016;63(3–4):248–55. doi: 10.2152/jmi.63.248
A118. Moon M, Lee SK. Applying of decision tree analysis to risk factors associated with pressure ulcers in long-term care facilities. Healthc Inform Res. 2017;23(1):43–52. doi: 10.4258/hir.2017.23.1.43
A119. Kaewprag P, et al. Predictive models for pressure ulcers from intensive care unit electronic health records using Bayesian networks. BMC Med Inform Decis Mak. 2017;17:81–91. doi: 10.1186/s12911-017-0471-z
A120. Deng X, Yu T, Hu A. Predicting the risk for hospital-acquired pressure ulcers in critical care patients. Crit Care Nurse. 2017;37(4):e1–11. doi: 10.4037/ccn2017548
A121. Chen HL, et al. Artificial neural network: a method for prediction of surgery-related pressure injury in cardiovascular surgical patients. J Wound Ostomy Contin Nurs. 2018;45(1):26–30. doi: 10.1097/WON.0000000000000388
A122. Li HL, Lin SW, Hwang YT. Using nursing information and data mining to explore the factors that predict pressure injuries for patients at the end of life. CIN: Comput Inform Nurs. 2019;37(3):133–41. doi: 10.1097/CIN.0000000000000489
A123. Park SK, Park HA, Hwang H. Development and comparison of predictive models for pressure injuries in surgical patients: a retrospective case-control study. J Wound Ostomy Contin Nurs. 2019;46(4):291–7. doi: 10.1097/WON.0000000000000544
A124. Cai JY, et al. Predicting the development of surgery-related pressure injury using a machine learning algorithm model. J Nurs Res. 2021;29(1):e135. doi: 10.1097/JNR.0000000000000411
A125. Hyun S, et al. Prediction model for hospital-acquired pressure ulcer development: retrospective cohort study. JMIR Med Inform. 2019;7(3):e13785. doi: 10.2196/13785
Appendix 2.
List of studies excluded in the systematic review
E1. Tzeng HM. Forecasting: adopting the methodology of support vector machines to nursing research. J Nurs Res. 2006;14(2):154-60.
E2. Meyfroidt G, Güiza F, Cottem D, De Becker W, Van Loon K, Aerts JM, et al. Computerized prediction of intensive care unit discharge after cardiac surgery: development and validation of a Gaussian processes model. Crit Care Med. 2011;39(6):1339-45.
E3. Roederer A, Holmes JH, Lee I, Park S. Classification of vasospasm after aneurysmal subarachnoid hemorrhage using data-driven machine learning techniques. AMIA Annu Symp Proc. 2013;2013:1203-10.
E4. Somanchi S, Adhikari S, Lin A, Eneva S, Ghani R. Early prediction of cardiac arrest (Code Blue) using electronic medical records. In: Proceedings of the 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: Association for Computing Machinery; 2015. p. 2119-28.
E5. Brennan CW, Meng F, Meterko MM, D'Avolio LW. Feasibility of automating patient acuity measurement using a machine learning algorithm. J Nurs Adm. 2016;46(9):431-7.
E6. Elbattah M, Molloy O. Coupling simulation with machine learning: a hybrid approach for elderly discharge planning. In: Proceedings of the 2016 Winter Simulation Conference. Piscataway (NJ): IEEE; 2016. p. 1598-609.
E7. Inoue S, Ueda N, Nohara Y, Nakashima N. Recognizing and understanding nursing activities for a whole day with a big dataset. In: Proceedings of the 6th International Workshop on Human-centric sensing, networking, and systems. New York: Association for Computing Machinery; 2016. p. 13-8.
E8. Jung In P. Developing a predictive model for hospital-acquired catheter-associated urinary tract infections using electronic health records and nurse staffing data. Comput Inform Nurs. 2016;34(9):394-402.
E9. Christiansen DN, Olling K, Wee L. Including specific symptoms in clinical scoring: predictive modelling and nursing of swallowing pain. Stud Health Technol Inform. 2016;225:882-3.
E10. Fernandes DL, Siqueira-Batista R, Gomes AP, Souza CR, Da Costa IT, Cardoso FDS, et al. Investigation of the visual attention role in clinical bioethics decision-making using machine learning algorithms. Patient Prefer Adherence. 2017;11:747-56.
E11. Fritz RL, Cook DJ. Identifying varying health states in smart home sensor data: an expert-guided approach. IEEE J Biomed Health Inform. 2017;21(2):419-25.
E12. Shi M, Steenhard D, Dong Y, Horsley S, Prewitt T, Weidenborner S, et al. A predictive model to identify individuals with diabetes at high risk for developing foot wounds using administrative data and medical records. Diabetes Care. 2017;40(12):1688-95.
E13. Steenhard D, Wei Y, Dong Y, Andrews G, Gopal V. A prediction model to identify individuals at high risk for developing heart failure using administrative data and medical records. J Am Heart Assoc. 2017;6(5):e005081.
E14. Wellner B, Grand J, Canzone E, Coarr M, Brady PW, Simmons J, et al. Predicting unplanned transfers to the intensive care unit: a machine learning approach leveraging diverse clinical elements. JMIR Med Inform. 2017;5(4):e46.
E15. Choi JH, Park HK, Park JE, Lee CM, Choi BK. [Prediction of turnover of new nurses using artificial intelligence]. J Korean Acad Nurs Adm. 2018;24(5):432-41. Korean.
E16. Blanchard TC, Willetts J, O'Connell MR, Chaudhuri S, Usvyat LA, Ellison BC, et al. A machine learning model to predict patient risk of peritonitis episodes. Kidney Med. 2019;1(4):175-83.
E17. Danesh V, Jones TL. Technology-enhanced surveillance: can facial expression analysis add predictive power to early warning scores? J Nurs Scholarsh. 2018;50(5):546-54.
E18. Dudding KM. Recognition of pain in the neonate to increase effective neonate to nurse communication. J Perinat Neonatal Nurs. 2018;32(4):341-8.
E19. Liu C, Hoi Q, Xing R, Kay TKT, Cheng SC, Si D. Computational psychiatric nursing research: scaling up the prediction of psychosis by natural language processing. Issues Ment Health Nurs. 2018;39(12):1044-51.
E20. Rojas JC, Venable LR, Fahrenbach JP, Carey KA, Edelson DP, Howell MD, et al. Predicting hospital length of stay after intensive care unit discharge with machine learning. Ann Am Thorac Soc. 2018;15(11):1286-93.
E21. Tran N, Lee J. Using multiple sentiment dimensions of nursing notes to predict mortality in the intensive care unit. J Biomed Inform. 2018;85:178-86.
E22. Veeranki SPK, Hayn D, Kramer D, Jauk S, Schreier G. Effect of nursing assessment on predictive delirium models in hospitalised patients. Stud Health Technol Inform. 2018;248:301-8.
E23. Zhao CY, Xu-Wilson M, Gangireddy SR, Horng S. Predicting disposition decision, mortality, and readmission for acute heart failure patients in the emergency department using vital sign, laboratory, echocardiographic, and other clinical data. J Am Med Inform Assoc. 2018;25(11):1443-52.
E24. Cho I, Jin I. Responses of staff nurses to an EMR-based clinical decision support service for predicting inpatient fall risk. Stud Health Technol Inform. 2019;264:433-7.
E25. Choi HA, Savarraj JPJ, Hergenroeder G, Zhu L, Chang T, Park S, et al. Machine learning improves the prediction of delayed cerebral ischemia and functional outcomes after subarachnoid hemorrhage. J Neurosurg. 2019;132(6):1940-8.
E26. Cramer EM, Seneviratne MG, Sharifi H, Ozturk A, Hernandez-Boussard T. Predicting the incidence of pressure ulcers in the intensive care unit using machine learning. AMIA Annu Symp Proc. 2019;2019:323-31.
E27. Frisch SO, Li H, Faramand Z, Martin-Gill C, Callaway CW, Sejdic E, et al. Using predictive machine learning modeling for the nursing triage of acute chest pain at the emergency department. Nurs Res. 2019;68(5):385-94.
E28. Gomes DC, Oliveira LES, Cubas MR, Barra CMCM. Use of computational tools as support to the cross-mapping method between clinical terminologies. Rev Lat Am Enfermagem. 2019;27:e3184.
E29. Gregoire JM, Subramanian N, Papazian D, Bersini H. Screening atrial fibrillation using machine learning. Comput Cardiol (2010). 2019;46:1-4.
E30. Cho I, Jin I. Responses of staff nurses to an EMR-based clinical decision support service for predicting inpatient fall risk. Stud Health Technol Inform. 2019;264:433-7.
E31. Kemp K, D'Souza A, Quan H, Santana M. A machine learning approach to predict risk of 30-day readmission: insights from hospital experience surveys completed by patients living with chronic conditions. BMC Med Inform Decis Mak. 2019;19(1):227.
E32. Koyner JL, Afshar M, Gilbert ER, Carey K, Churpek MM. External validation of an electronic health record (EHR)-based machine learning risk score for hospital-based AKI. Clin J Am Soc Nephrol. 2019;14(10):1448-55.
E33. Rockwood K, Shehzad A, Stanley J, Dunn T, Howlett S, Mitnitski A, et al. Development of a machine learning algorithm to classify dementia stage based on symptoms reported online. J Alzheimers Dis. 2019;69(1):177-85.
E34. Royer J, Signorovitch J, Pivneva I, Huber W, Capkun G. PSY24 Early predictors of Sjögren's syndrome: a machine learning approach. Ann Rheum Dis. 2019;78(Suppl 2):2098.
E35. Topaz M, Murga L, Bar-Bachar O, Cato K, Collins S. Extracting alcohol and substance abuse status from clinical notes: the added value of nursing data. Comput Inform Nurs. 2019;37(3):124-30.
E36. Zampieri FG, Salluh JIF, Azevedo LCP, Kahn JM, Damiani LP, Borges LP, et al. ICU staffing feature phenotypes and their relationship with patients’ outcomes: an unsupervised machine learning analysis. Intensive Care Med. 2019;45(11):1599-607.
E37. Zarei K, Shinozaki G. Prediction of delirium, mortality, and fall risk in inpatients using bispectral EEG. Gen Hosp Psychiatry. 2019;57:17-23.
E38. Coombs LA, Orlando A, Adamson BJS, Griffith SD, Lakhtakia S, Rich A, et al. Prospective validation of a clinical tool developed with machine learning to identify high-risk patients with cancer and reduce emergency department visits. JCO Clin Cancer Inform. 2020;4:462-72.
E39. Faisal AAM, Siraj MS, Abdullah MT, Shahid O, Abir FF, Ahad MAR. A pragmatic signal processing approach for nurse care activity recognition using classical machine learning. In: 2020 International Conference on Signal Processing and Information Security (ICSPIS). Piscataway (NJ): IEEE; 2020. p. 1-6.
E40. Fralick M, Dai D, Pou-Prom C, Verma AA, Mamdani M. Identifying adults at risk of unintentional severe hypoglycemia in hospital using artificial intelligence (RUSHH-AI). BMJ Health Care Inform. 2020;27(1):e100115.
E41. Griner TE, Thompson M, High H, Buckles J. Artificial intelligence forecasting census and supporting early decisions. Nurs Adm Q. 2020;44(3):250-6.
E42. Kang Y, Hurdle J. Predictive model for risk of 30-day rehospitalization using a natural language processing/machine learning approach among Medicare patients with heart failure. J Cardiovasc Nurs. 2020;35(4):374-82.
E43. Li Q, Su Q, Lin Y, Deng G. Pressure injury analysis and prediction based on machine learning methods. In: Proceedings of the 2020 4th International Conference on Electronic Information Technology and Computer Engineering. New York: Association for Computing Machinery; 2020. p. 770-4.
E44. Moen H, Hakala K, Peltonen LM, Suhonen H, Ginter F, Salakoski T, et al. Supporting the use of standardized nursing terminologies with automatic subject heading prediction: a comparison of sentence-level text classification methods. J Biomed Inform. 2020;105:103407.
E45. Noh YJ, Lee JH. Classification of delirium patients using machine learning technology. Int J Environ Res Public Health. 2020;17(20):7398.
E46. Phan T, Srikanth V, Cadilhac D, Kim J, Bladin C, Dewey H, et al. Machine learning of future framingham risk score in standirm trial. J Clin Hypertens (Greenwich). 2020;22(12):2205-11.
E47. Pierce L, Shah SJ. Maintaining mobility: predicting mobility loss during hospitalization on hospital day one. Gerontol Geriatr Med. 2020;6:2333721420919195.
E48. Solberg LM, Ingibjargardottir R, Wu Y, Lucero R. Nursing innovations in machine learning: using natural language processing in falls prediction. J Gerontol Nurs. 2020;46(8):13-8.
E49. Zachariah P, Sanabria E, Liu J, Cohen B, Yao D, Larson E. Novel strategies for predicting healthcare-associated infections at admission: implications for nursing care. J Nurs Adm. 2020;50(3):139-45.
E50. Zhan T, Goyal D, Guttag J, Mehta R, Elahi Z, Syed Z, et al. Machine intelligence for early targeted precision management and response to outbreaks of respiratory infections. AMIA Annu Symp Proc. 2020;2020:1367-76.
E51. Afnan MAM, Ali F, Worthington H, Netke T, Singh P, Kajamuhan C. Triage nurse prediction as a covariate in a machine learning prediction algorithm for hospital admission from the emergency department. In: Proceedings of the 2021 International Conference on Digital Health. New York: Association for Computing Machinery; 2021. p. 89-98.
E52. Aoki M, Yokota S, Kagawa R, Shinohara E, Imai T, Ohe K. Automatic classification of electronic nursing narrative records based on Japanese standard terminology for nursing. Stud Health Technol Inform. 2021;281:785-9.
E53. Cho I, Jin IS, Park H, Dykes PC. Clinical impact of an analytic tool for predicting the fall risk in inpatients: controlled interrupted time series. J Med Internet Res. 2021;23(9):e27989.
E54. De Silva K, Mathews N, Teede H, Forbes A, Jönsson D, Demmer RT, et al. Clinical notes as prognostic markers of mortality associated with diabetes mellitus following critical care: a retrospective cohort analysis using machine learning and unstructured big data. Diabet Med. 2021;38(11):e14674.
E55. Duprey MS, Capuano A, Lee Y, Daiello L, Kiel DP, Kim D, et al. Model development to predict hip fracture in nursing home residents. J Am Geriatr Soc. 2021;69(11):3237-45.
E56. Ganju A, Satyan S, Tanna V, Menezes SR. AI for improving children’s health: a community case study. In: Proceedings of the 2021 International Conference on Information Technology. New York: Association for Computing Machinery; 2021. p. 45-51.
E57. Gong K, Lu R, Bergamaschi T, Sanyal A, Guo J, Kim H, et al. Decoding digital health signatures for prediction of delirium in the intensive care unit. In: Proceedings of the 2nd International Workshop on Health Intelligence. New York: Association for Computing Machinery; 2021. p. 11-20.
E58. Jago R, van der Gaag A, Stathis K, Petej I, Lertvittayakumjorn P, Krishnamurthy Y, et al. Use of artificial intelligence in regulatory decision-making. Clin Pharmacol Ther. 2021;110(3):616-20.
E59. Ocagli H, Lorenzoni G, Bottigliengo D, Azzolina D, Stivanello L, Giorato E, et al. The SYSTEMIC project: a pilot study to develop a registry of patients with an ostomy for predictive modeling of outcomes. J Wound Ostomy Continence Nurs. 2021;48(5):401-6.
E60. Saha P, Sircar R, Bose A. Using hospital Admission, Discharge & Transfer (ADT) data for predicting readmissions. In: 2021 IEEE International Conference on Big Data (Big Data). Piscataway (NJ): IEEE; 2021. p. 3020-5.
E61. Saleh M, Abbas M, Prud'homme J, Somme D, Le Bouquin Jeannes R. A reliable fall detection system based on analyzing the physical activities of older adults living in long-term care facilities. Sensors (Basel). 2021;21(4):1199.
E62. Sayem FR, Sheikh MDM, Ahad MAR. Feature-based method for nurse care complex activity recognition from accelerometer sensor. In: 2021 International Conference on Electronics, Communications and Information Technology (ICECIT). Piscataway (NJ): IEEE; 2021. p. 1-4.
E63. Scheets P, Billings M, Dhamija K, Hennessy P. Physical performance is significantly more important than health condition when considering rehabilitation outcome. J Geriatr Phys Ther. 2021;44(3):142-7.
E64. Song J, Gao Y, Yin P, Li Y, Li Y, Zhang J, et al. The random forest model has the best accuracy among the four pressure ulcer prediction models using machine learning algorithms. Int Wound J. 2021;18(6):880-9.
E65. Thepmankorn P, Heshmati K, Souayah S, Shafiq B, Adam T, Adam N, et al. Effect of neurological manifestations on SARS-CoV-2 infection prognosis using machine learning models. Front Neurol. 2021;12:699298.
E66. Tran S, Bunney G, Han S, Wang H, Gu C, Luo Y, et al. 58 Using machine learning to predict hospital disposition with geriatric emergency department innovations intervention. Ann Emerg Med. 2021;78(4 Suppl):S24-5.
E67. Wang P, Luo Z, Guo Z, Li D, Wang Y. Machine learning-based models for prediction of nursing staff mental health status. Front Psychiatry. 2021;12:732958.
E68. Cohen B, Sanabria E, Liu J, Zachariah P, Shang J, Song J, et al. Predicting healthcare-associated infections, length of stay, and mortality with the nursing intensity of care index. J Am Med Inform Assoc. 2022;29(2):281-9.
E69. Depauw L, Bogaert S, De Corte T, Vermassen J, Colpaert K, Decruyenaere J. PREMIUMS: predicting mortality in ICU patients by healthcare workers, scoring systems and artificial intelligence. Crit Care. 2022;26(1):164.
E70. Durieux BN, Tarbi EC, Lindvall C. Opportunities for computational tools in palliative care: supporting patient needs and lowering burden. Curr Opin Support Palliat Care. 2022;16(2):81-7.
E71. Fralick M, Debnath M, Pou-Prom C, O’Brien P, Perkins BA, Carson E, et al. Using real-time machine learning to prevent in-hospital severe hypoglycemia: a prospective study. Diabetes Care. 2022;45(8):1797-802.
E72. Garriga R, Mas J, Abraha S, Nolan J, Harrison O, Tadros G, et al. Machine learning model to predict mental health crises from electronic health records. Nat Med. 2022;28(7):1435-42.
E73. Kang Y, Topaz M, Dunbar SB, Stehlik J, Hurdle J. The utility of nursing notes among Medicare patients with heart failure to predict 30-day rehospitalization: a pilot study. Comput Inform Nurs. 2022;40(6):392-401.
E74. Li FW, Li P, Bai CH. Development of a depression prediction model for renal transplant recipients based on machine learning. BMC Psychiatry. 2022;22(1):475.
E75. Reunamo A, Peltonen LM, Mustonen R, Saari M, Salakoski T, Salanterä S, et al. Text classification model explainability for keyword extraction - towards keyword-based summarization of nursing care episodes. Stud Health Technol Inform. 2022;289:33-6.
E76. Shim S, Yu JY, Jekal S, Song YJ, Moon KT, Lee JH, et al. Development and validation of interpretable machine learning models for inpatient fall events and electronic medical record integration. Int J Med Inform. 2022;167:104868.
E77. Son CS, Kang WS, Lee JH, Moon KJ. Machine learning to identify psychomotor behaviors of delirium for patients in long-term care facility. Int J Environ Res Public Health. 2022;19(18):11603.
E78. Yu F, Zhang X, Gao L, Lv J, Lin X. Association between tongue muscle quality and swallowing disorders in older nursing home residents. Geriatr Nurs. 2022;48:153-7.
E79. Alkhalaf M, Mengyang YIN, Chao D, Chang HC, Yu P. Machine learning model to extract malnutrition data from nursing notes. Stud Health Technol Inform. 2023;305:273-7.
E80. Ambushe SA, Awoke N, Demissie BW, Tekalign T. Holistic nursing care practice and associated factors among nurses in public hospitals of Wolaita zone, South Ethiopia. SAGE Open Nurs. 2023;9:23779608231165448.
E81. Branum C, Schiavenato M. Can ChatGPT accurately answer a PICOT question? assessing AI response to a clinical question. Nurse Educ. 2023;48(5):253-6.
E82. Chae S, Davoudi A, Song J, Evans L, Hobensack M, Bowles KH, et al. Predicting emergency department visits and hospitalizations for patients with heart failure in home healthcare using a time series risk model. J Am Med Inform Assoc. 2023;30(10):1676-86.
E83. Cho I, Kim M, Song MR, Dykes PC. Evaluation of an approach to clinical decision support for preventing inpatient falls: a pragmatic trial. J Am Med Inform Assoc. 2023;30(8):1378-85.
E84. Dereli O, Schramm A, Behringer J, Berthele A, Hapfelmeier A, Hemmer B, et al. Identifying informative factors in assessment of disease progression for multiple sclerosis using healthcare records. Mult Scler. 2023;29(4-5):565-74.
E85. Duan J, Li H, Ma X, Zhang H, Lasky R, Monaghan CK, et al. Predicting SARS-CoV-2 infection among hemodialysis patients using multimodal data. NDT Plus. 2023;16(5):982-90.
E86. Giesa N, Haufe S, Menk M, Weiß B, Spies C, Piper SK, et al. Predicting postoperative delirium assessed by the nursing screening delirium scale in the recovery room for non-cardiac surgeries without craniotomy: a retrospective study using a machine learning approach. J Clin Med. 2023;12(5):2032.
E87. Govindan S, Spicer A, Bearce M, Martin J, Karway G, Schaefer R, et al. Development and external validation of a national veterans COVID-19 machine learning model. Am J Respir Crit Care Med. 2023;207(4):442-51.
E88. Huang S, Teng Y, Du JJ, Zhou X, Duan F, Feng C. Internal and external validation of machine learning-assisted prediction models for mechanical ventilation-associated severe acute kidney injury. Front Med (Lausanne). 2023;10:1107297.
E89. Kim TH, Kim M, Kim D, Heo JW, Moon D, Heo Y, et al. Development of a predictive model for the occurrence of pressure ulcers of patients in intensive care units using AI. Sci Rep. 2023;13(1):3192.
E90. Lei L, Zhang S, Yan L, Yang C, Liu ZQ, Xu H, et al. Machine learning-based prediction of delirium 24 h after pediatric intensive care unit admission in critically ill children: a prospective cohort study. BMC Med Inform Decis Mak. 2023;23(1):153.
E91. Li F, Tao Z, Li R, Qu Z. The early warning research on nursing care of stroke patients with intelligent wearable devices under COVID-19. Am J Transl Res. 2023;15(4):2875-81.
E92. Melo M, Brandão-de-Resende C, Day A, Lee E, Neo YN, Jindal A. Predicting risk stratification in ophthalmology emergency department triage from patient-informed structured clinical features using machine learning: preliminary results. BMJ Health Care Inform. 2023;30(1):e100732.
E93. Park M, Moon KJ. Web-based delirium prevention application for long-term care facilities. Healthc Inform Res. 2023;29(1):72-8.
E94. Pouzols S, Despraz J, Mabire C, Raisaro JL. Development of a predictive model for hospital-acquired pressure injuries. Comput Methods Programs Biomed. 2023;241:107759.
E95. Saputra DCE, Sunat K, Ratnaningsih T. A new artificial intelligence approach using extreme learning machine as the potentially effective model to predict and analyze the diagnosis of anemia. Inform Med Unlocked. 2023;37:101185.
E96. Sjövall F, Persson I. A machine learning algorithm for early prediction of sepsis in intensive care. J Clin Med. 2023;12(13):4399.
E97. Tehrany PM, Zabihi MR, Ghorbani Vajargah P, Tamimi P, Ghaderi A, Norouzkhani N, et al. Risk predictions of hospital-acquired pressure injury in the intensive care unit based on a machine learning algorithm. Wounds. 2023;35(7):151-5.
E98. To D, Carey K, Joyce C, Salisbury-Afshar E, Edelson DP, Churpek MM, et al. Predicting intensive care unit admissions in patients with alcohol withdrawal using electronic health record data. Alcohol Clin Exp Res. 2023;47(5):981-91.
E99. Wang J, Chen H, Wang H, Liu W, Peng D, Zhao Q, et al. A risk prediction model for physical restraints among older Chinese adults in long-term care facilities: machine learning study. BMC Geriatr. 2023;23(1):162.
E100. Wilson PM, Ramar P, Philpot LM, Soleimani J, Ebbert JO, Storlie CB, et al. Effect of an artificial intelligence decision support tool on palliative care referral in hospitalized patients: a randomized clinical trial. JAMA Netw Open. 2023;6(7):e2324451.
E101. Wu EL, Wu CY, Lee MB, Chu KC, Huang MS. Development of internet suicide message identification and the monitoring-tracking-rescuing model in Taiwan. J Formos Med Assoc. 2023;122(1 Pt 1):44-53.
E102. TUTORIAL. 2024.
E103. Alkhalaf M, Shen J, Chang HC, Deng C, Yu P. Fine-tuning large language models for effective nutrition support in residential aged care: a domain expertise approach. J Am Med Inform Assoc. 2024;31(5):1098-106.
E104. Ben-Sasson A, Guedalia J, Ilan K, Shaham M, Shefer G, Cohen R, et al. Predicting autism traits from baby wellness records: a machine learning approach. J Autism Dev Disord. 2024;54(1):210-23.
E105. Bhimani AA, Frenkel TS, Hasham AK. Can artificial intelligence be utilized to predict real-time adverse outcomes in individuals arriving at the emergency department with hyperglycemic crises?: implications for APRN practice. J Emerg Nurs. 2024;50(1):132-7.
E106. Brann F, Sterling NW, Frisch SO, Schrager JD. Sepsis prediction at emergency department triage using natural language processing: retrospective cohort study. J Med Internet Res. 2024;26:e52467.
E107. Conway A, Li J, Rad MG, Mafeld S, Taati B. Automating sedation state assessments using natural language processing. Intensive Crit Care Nurs. 2024;80:103565.
E108. Doran K, Witmer S, Yoon KL, Fischer ER, Ebangwese A, Sharma S, et al. Gauging the stress of long-term care nursing assistants using ecological momentary assessment, wearable sensors and end of day reconstruction. J Adv Nurs. 2024;80(5):2068-80.
E109. Doshi H, Deshpande K. Burden of fever and hospital mortality in patients admitted to the intensive care unit with isolated traumatic brain injury-a retrospective cohort study using continuous temperature data. J Neurotrauma. 2024;41(7-8):938-45.
E110. Hu H, Hong S, Jia Y, Song J. Research progress on application of machine learning in discharge preparation service for patients. Chin Nurs Res. 2024;11:100065.
E111. Hunstein D, Fiebig M. Staff management with AI: predicting the nursing workload. Stud Health Technol Inform. 2024;316:16-24.
E112. Hunstein D, Frischen L, Fiebig M. Development of a data model to predict nursing workload using routine clinical data. Stud Health Technol Inform. 2024;316:9-15.
E113. Jauk S, Kramer D, Sumerauer S, Veeranki SPK, Schrempf M, Puchwein P. Machine learning-based delirium prediction in surgical in-patients: a prospective validation study. Wien Klin Wochenschr. 2024;136(3-4):81-8.
E114. Jeffery AD, Fabbri D, Reeves RM, Matheny ME. Use of noisy labels as weak learners to identify incompletely ascertainable outcomes: a feasibility study with opioid-induced respiratory depression. J Am Med Inform Assoc. 2024;31(1):50-7.
E115. Khalil CABI, Saab A, Rahme J, Abla J, Seroussi B. Evaluation of machine learning algorithms for pressure injury risk assessment in a hospital with limited IT resources. Stud Health Technol Inform. 2024;336:205-12.
E116. Kim H, Park H, Kang S, Kim J, Kim J, Jung J, et al. Evaluating the validity of the nursing statements algorithmically generated based on the International Classifications of Nursing Practice for respiratory nursing care using large language models. Int J Med Inform. 2024;185:105378.
E117. Liu Q, Tong G, Ye Q. Analysis of frailty in peritoneal dialysis patients based on logistic regression model and XGBoost model. Ren Fail. 2024;46(1):2303036.
E118. Liu T, Chen Z, Wu Y, Guo Y, Chatzkel JA, Bian J. Clinical large language model to predict loss to follow up for oncology patients discharged to skilled nursing facilities. JAMIA Open. 2024;7(1):ooae013.
E119. Lukkahatai N, Han G. Perspectives on artificial intelligence in nursing in Asia. Nurs Outlook. 2024;72(2):102143.
E120. Meeus M, Beirnaert C, Mahieu L, Laukens K, Meysman P, Mulder A, et al. Clinical decision support for improved neonatal care: the development of a machine learning model for the prediction of late-onset sepsis and necrotizing enterocolitis. Eur J Pediatr. 2024;183(2):643-52.
E121. Özsezer G, Mermer G. Prediction of drinking water quality with machine learning models: a public health nursing approach. Public Health Nurs. 2024;41(1):108-15.
E122. Patel M, Mara A, Acker Y, Gollon J, Wolf S, Zafar Y, et al. Using machine learning for targeted advance care planning (ACP) conversations in patients with cancer: a quality improvement initiative. JCO Oncol Pract. 2024;20(2):223-31.
E123. Ponce-Valencia A, Jiménez-Rodríguez D, Hernández Morante JJ, Martínez Cortés C, Pérez-Sánchez H, Echevarría Pérez P. An interpretable machine learning approach to predict sensory processing sensitivity trait in nursing students. Comput Inform Nurs. 2024;42(2):137-45.
E124. Su C, Huang L, Zhong J, Peng L, Wang J, Gao H, et al. The machine learning algorithm screened the characteristic variables of prolonged hospital stay after hip fracture and constructed the prediction model. Front Public Health. 2024;11:1309322.
E125. Sung S, Kim Y, Kim SH, Jung H. Identification of predictors for clinical deterioration in patients with COVID-19 via electronic nursing records: retrospective observational study. Healthc Inform Res. 2024;30(1):64-74.
E126. Tiase VL, Sward KA, Facelli J. Developing a conceptual data model for nursing workload. Nurs Outlook. 2024;72(2):102148.
E127. Tiase VL, Sward KA, Facelli JC. A scalable and extensible logical data model of electronic health record audit logs for temporal data mining (RNteract): model conceptualization and formulation. J Am Med Inform Assoc. 2024;31(2):330-9.
E128. To D, Steel TL, Carey K, Joyce C, Salisbury-Afshar E, Mayampurath A, et al. Natural language processing and machine learning to predict high-intensity care in patients with alcohol withdrawal syndrome. NPJ Digit Med. 2024;7(1):50.
E129. Tobias PF, Oliver Z, Huang Y, Bayne C, Fidyk L. Using acuity to predict oncology infusion center daily nurse staffing and outcomes. J Nurs Adm. 2024;54(2):80-7.
E130. Wang Y, Wu H. Research on intelligent management of nursing quality based on decision tree algorithm. In: 2024 3rd International Conference on Big Data, Information and Computer Network (BDICN). Piscataway (NJ): IEEE; 2024. p. 306-11.
E131. Wen Z, Wang Y, Chen S, Li Y, Deng H, Pang H, et al. Construction of a predictive model for postoperative hospitalization time in colorectal cancer patients based on interpretable machine learning algorithm: a prospective preliminary study. Front Public Health. 2024;12:1274191.
E132. White K, Marzolf S, Gore T. Incorporating an Alzheimer's unfolding case study for health profession student engagement. J Allied Health. 2024;53(1):e53-7.
E133. Windle N, Alam A, Patel H, Street JM, Lathwood M, Farrington T, et al. Predicting hospitalization risk among home care residents in the United Kingdom: development and validation of a machine learning-based predictive model. J Med Internet Res. 2024;26:e52466.
E134. Wynn MO, Goldstone L, Gupta R, Allport J, Fraser RDJ. Improving pressure injury risk assessment using real-world data from skilled nursing facilities: a cohort study. J Am Med Dir Assoc. 2024;25(1):144-50.e3.
E135. Xia Q, Huang Q, Li J, Xu Y, Ge S, Zhang X, et al. Evaluating the quality of home care in community health service centres: a machine learning approach. J Nurs Manag. 2024;2024:8840212.
E136. Yan Z, Quan G, Jia-Hui X. Criticality of nursing care for patients with Alzheimer's disease in the ICU: insights from MIMIC III dataset. Geriatr Nurs. 2024;56:123-30.
E137. Yang PC, Jha A, Xu W, Song Z, Jamp P, Teuteberg JJ. Cloud-based machine learning platform to predict clinical outcomes at home for patients with cardiovascular conditions discharged from hospital: clinical trial. JMIR Cardio. 2024;8:e52003.
E138. Yang X, Li YM, Wang Q, Li R, Zhang P. Machine learning model based on RCA-PDCA nursing methods and differentiating factors to predict hypotension during cesarean section surgery. J Healthc Eng. 2024;2024:1905204.
E139. Zhang H, Ouyang Y, Huang X. Prediction model of uroschesis rate after radical cervical cancer resection based on machine learning. Medicine (Baltimore). 2024;103(6):e37033.
E140. Zhang L, Yu R, Chen K, Zhang Y, Li Q, Chen Y. Enhancing deep vein thrombosis prediction in patients with coronavirus disease 2019 using improved machine learning model. J Vasc Surg Venous Lymphat Disord. 2024;12(1):101824.
E141. Zhang S, Wang J, Zhang B, Liu X. Construction and validation of a prediction model of aspiration risk of acute poisoning patients during gastric lavage. BMC Emerg Med. 2024;24(1):4.
E142. Guilamet MCV, Bernauer M, Micek ST, Kollef MH. Cluster analysis to define distinct clinical phenotypes among septic patients with bloodstream infections. PLoS One. 2019;14(6):e0217711.
E143. Romero Valencia S, Gomez Cantarino S, Sim-Sim M, Espina B, Mendes D. Automatic learning for improvement in joint mobility in the elderly. J Healthc Eng. 2019;2019:4213962.
E144. Choi JH, Jeong MD. [Ensemble model based on mixed machine learning for urine spectrum analysis]. J Korea Inst Inf Commun Eng. 2020;24(12):1668-76. Korean.
E145. An C, Lim H, Kim DW, Chang JH, Choi YJ, Kim SW. Machine learning prediction for mortality of patients diagnosed with COVID-19: a nationwide Korean cohort study. Sci Rep. 2020;10(1):18731.
E146. Jeon E, Kim Y, Park H, Park RW, Shin H, Park HA. Analysis of adverse drug reactions identified in nursing notes using reinforcement learning. Healthc Inform Res. 2020;26(4):303-12.
E147. Korach ZT, Yang S, Rossetti SC, Cato KD, Kang MJ, Knaplund C, et al. Mining clinical phrases from nursing notes to discover risk factors of patient deterioration. Int J Med Inform. 2020;141:104230.
E148. Mairittha T, Mairittha N, Inoue S. Automatic labeled dialogue generation for nursing record systems. In: Proceedings of the 12th Language Resources and Evaluation Conference. Marseille: European Language Resources Association; 2020. p. 4363-9.
E149. Abbott EE, Oh W, Dai Y, Feuer L, Chan LL, Carr G, et al. Joint modeling of social determinants and clinical factors to define subphenotypes in out-of-hospital cardiac arrest survival: cluster analysis. J Med Internet Res. 2023;25:e44806.
E150. Arjun VS, Chandrasekhar L, Jaseena KU. Wound tissue segmentation and classification using U-net and random forest. Int J Cogn Comput Eng. 2024;5:59-69.
E151. Hayashi T, Cimr D, Studnicka F, Fujita H, Busovsky D, Cimler R. Patient deterioration detection using one-class classification via cluster period estimation subtask. Artif Intell Med. 2024;148:102766.
E152. Luo CH, Mao B, Wu Y, He Y. The research hotspots and theme trends of artificial intelligence in nurse education: a bibliometric analysis from 1994 to 2023. Nurse Educ Today. 2024;135:106132.

Figure & Data

References

    Citations

    Citations to this article as recorded by  

      Download Citation

      Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

      Format:

      Include:

      Machine Learning Applications in Nursing-Affiliated Research: A Systematic Review
      Korean J Adult Nurs. 2025;37(3):189-214.   Published online August 29, 2025
      DOI: https://doi.org/10.7475/kjan.2025.0327
      Download Citation
      Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.
      Format:
      • RIS — For EndNote, ProCite, RefWorks, and most other reference management software
      • BibTeX — For JabRef, BibDesk, and other BibTeX-specific software
      Include:
      • Citation for the content below
      Machine Learning Applications in Nursing-Affiliated Research: A Systematic Review
      Korean J Adult Nurs. 2025;37(3):189-214.   Published online August 29, 2025
      DOI: https://doi.org/10.7475/kjan.2025.0327
      Close

      Figure

      • 0

      Related articles

      Machine Learning Applications in Nursing-Affiliated Research: A Systematic Review
      Image
      Figure 1. PRISMA flow diagram.

      Figure 1.

      Machine Learning Applications in Nursing-Affiliated Research: A Systematic Review
      No. Phases Short descriptions
      1 Business understanding - Determine business objectives, assess situation, define data mining goals, develop project plan.
      Understanding business goals and translating them into data mining objectives.
      2 Data understanding - Initial data collection, data description, data exploration, data quality assessment.
      Collecting data, familiarizing oneself with data, identifying quality issues, and gaining initial insights.
      3 Data preparation - Data selection, data cleaning, data construction, data integration, data formatting.
      All activities required to construct the final dataset from initial raw data.
      4 Modeling - Select modeling techniques, design tests, build models, evaluate models.
      Selecting, applying, and optimizing modeling techniques.
      5 Evaluation - Evaluate results, review processes, determine next steps.
      Evaluating the model from the perspective of achieving business objectives and reviewing the entire process.
      6 Deployment - Plan deployment, monitor and maintain the model, produce final reports, review project.
      Integrating the model into actual business processes and ensuring organizational usage of the outcomes.
      No. Phases Data extraction details
      1 General characteristics 1) Journal name, 2) Year of publication, 3) Authors, 4) Proportion of nursing-affiliated authors (determined based on institutional affiliation), 5) Country, 6) Type of ML (supervised/unsupervised/reinforcement), 7) Type of algorithm used (prediction/classification/clustering, etc.)
      2 Business understanding 8) Research objective, 9) Research design and methodology (KDD, CRISP-DM, etc.)
      3 Data understanding 10) Tools and software used for data analysis, 11) Data source, 12) EDA methods, 13) Target of ML application, 14) Sample size
      4 Data preparation 15) Data preprocessing techniques (normalization, standardization, encoding, etc.), 16) Predictor (explanatory/training) variables, 17) Number of predictor variables, 18) Target variable (the variable to be predicted, classified, or analyzed), 19) Data split ratio
      5 Modeling 20) Applied ML algorithms and modeling techniques, 21) Hyperparameter tuning methods
      6 Evaluation 22) Confusion matrix, 23) Performance evaluation metrics for classification and regression models, 24) Performance evaluation results for each model, 25) Best model, 26) Method of analyzing variable importance or the impact on the target variable (feature importance, SHAP value, etc.)
      7 Deployment 27) Research environment, 28) IRB approval, 29) Whether the model was actually deployed
      Characters Categories n (%)
      Location† Asia 71 (35.9)
      North America 59 (29.8)
      Europe 22 (11.1)
      Oceania 4 (2.0)
      South America 2 (1.0)
      Publication year Before 2020 26 (20.8)
      2020 11 (8.8)
      2021 20 (16.0)
      2022 18 (14.4)
      2023 24 (19.2)
      2024 26 (20.8)
      Percentage of nurses on the research team (%) 0.0–0.25 16 (12.8)
      0.25–0.5 26 (20.8)
      0.5–0.75 41 (32.8)
      0.75–1.0 42 (33.6)
      ML type Supervised learning 117 (93.6)
      Unsupervised learning 5 (4.0)
      Supervised and unsupervised learning 3 (2.4)
      Algorithm type Classification 101 (80.8)
      Regression 14 (11.2)
      Clustering 7 (5.6)
      Classification, association rule mining 2 (1.6)
      Regression, dimensionality reduction 1 (0.8)
      Research objective Pressure-injury/ulcer 24 (19.2)
      Readmission/utilization 17 (13.6)
      Fall risk 10 (8.0)
      Others 74 (59.2)
      Journal CIN: Computers, Informatics, Nursing 7 (5.6)
      Applied Clinical Informatics 4 (3.2)
      International Journal of Environmental Research and Public Health 4 (3.2)
      Journal of the American Medical Informatics Association 4 (3.2)
      BMC Medical Informatics and Decision Making 3 (2.4)
      International Journal of Medical Informatics 3 (2.4)
      Journal of Advanced Nursing 3 (2.4)
      Journal of Nursing Management 3 (2.4)
      Nursing Research 3 (2.4)
      BMC Nursing 2 (1.6)
      Journal of Biomedical Informatics 2 (1.6)
      Journal of Emergency Nursing 2 (1.6)
      Journal of Medical Internet Research 2 (1.6)
      Journal of Nursing Scholarship 2 (1.6)
      Journal of the American Medical Directors Association 2 (1.6)
      International Journal of Nursing Studies 2 (1.6)
      Healthcare 2 (1.6)
      Innovation in Applied Nursing Informatics 2 (1.6)
      Nurse Education Today 2 (1.6)
      Archives of Psychiatric Nursing 2 (1.6)
      Others 69 (55.2)
      Sections Key findings
      Quality appraisal of the studies The TRIPOD+AI appraisal showed moderate compliance (≈50%). Core methods were well reported (>85%), but transparency and ethical aspects were weak (<25%).
      General characteristics of the selected studies USA (30.8%), China (14.5%), and Korea (10.7%) dominated. Most studies used supervised learning (93.6%), especially classification tasks (80.8%).
      Problem definition for research objective Top topics were pressure injury, fall, and readmission. Most studies lacked formal data-mining frameworks; CRISP-DM was the most used among those that did.
      Data collection and exploration R, Python, and SPSS were the most used tools. EMR/EHR and survey data dominated. Sample sizes ranged from 5 to over 3.5 million cases.
      Data preparation Common preprocessing included standardization, normalization, and imputation. Label encoding and one-hot encoding were frequent. Most studies used <50 predictors.
      Model building Random forest was the most used algorithm, followed by logistic regression, SVM, and XGBoost. Hyperparameter tuning was often omitted; Grid Search was most common when used.
      Evaluation and review AUC-ROC, accuracy, sensitivity, and F1-score were most reported. RF was most often top-performing; 42 studies did not report variable importance.
      Deployment Hospitals were the most common setting. IRB approval was reported in 76.0% of studies. Only seven studies described actual deployment.
      Table 1. CRISP-DM Process Model Descriptions

      CRISP-DM=Cross-Industry Standard Process for Data Mining.

      Table 2. Data Extraction Plan

      CRISP-DM=Cross-Industry Standard Process for Data Mining; EDA=exploratory data analysis; IRB=Institutional Review Board; KDD=Knowledge Discovery in Databases; ML=machine learning; SHAP=Shapley additive explanations.

      Table 3. General Characteristics of the Selected Studies (N=125)

      ML=machine learning;

      The number of “location” values exceeds the 125 included studies because location data were compiled for each author, and a single study could involve authors from multiple locations.

      Table 4. Summary of the Findings of This Study

      AUC-ROC=area under the receiver operating characteristic curve; CRISP-DM=Cross-Industry Standard Process for Data Mining; EHR=electronic health record; EMR=electronic medical record; IRB=Institutional Review Board; RF=random forest; SVM=support vector machine; TRIPOD+AI=Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis+Artificial Intelligence extension; XGBoost=eXtreme gradient boosting.

      Table 1.

      Table 2.

      Table 3.

      Table 4.

      Korean J Adult Nurs : Korean Journal of Adult Nursing

      ABOUT

      BROWSE ARTICLES 

      EDITORIAL POLICY

      FOR CONTRIBUTORS  

      © Korean Society of Adult Nursing.
      TOP