Current State and Future Trends of Research on Pressure Injuries in Nursing Homes: Network Analysis and Topic Modeling
Article information
Abstract
Background
Pressure injuries (PIs) are a significant issue in nursing homes (NHs) that affect residents’ physical, psychological, and financial well-being. Despite advancements in prevention and management, PIs remain a critical challenge requiring a comprehensive analysis of research trends to guide future efforts.
Methods
We analyzed 1,857 research abstracts published up to December 2023 using text network analysis and topic modeling. Keyword frequency, co-occurrence patterns, and centrality metrics were examined to identify the key terms. Latent Dirichlet allocation was employed for topic modeling, and trends were analyzed over decades to explore the evolution of research priorities.
Results
The most frequent keywords were “patient,” “care,” and “wound,” emphasizing the central themes of patient care and wound management. Four main topics were identified: Managing PI Complications, Preventing PIs and Maintaining Skin Integrity, Improving Organizational Strategies for PI Care, and Managing Systemic Conditions in PI Care. Trends revealed a decline in studies on post-injury management, while prevention and organizational strategies have gained prominence. Patient-specific risk factors have consistently attracted attention over the decades, highlighting the importance of individualized care.
Conclusion
This study underscores the shift from reactive post-injury management to proactive prevention and organizational strategies in PI research. These findings emphasize the need for integrated approaches that combine patient-centered care with systemic and preventive measures to improve outcomes in NHs.
Introduction
The increasing older population has raised the demand for nursing homes (NHs) to support an aging society [1]. Residents of NHs, who often experience reduced mobility and prolonged exposure to pressure and shear forces, are at high risk of developing pressure injuries (PIs) [2]. Despite advancements in treatment methods and heightened national interest in preventing patient safety issues, PIs in NHs remain a significant challenge [3]. PIs cause severe physical and psychological pain in older adults, reduce their quality of life, and impose financial burdens on individuals, hospitals, and national healthcare systems [3-5]. Consequently, the prevention and management of PIs have become critical priorities in NHs and are reflected in national healthcare policies and institutional guidelines [6].
Given the adverse effects of PIs, numerous studies have explored their prevention and management in NHs [4,7-9]. However, there is a growing need to comprehensively analyze these studies to identify research trends and suggest future directions [10]. Systematic literature reviews and meta-analyses are commonly used to synthesize research findings [10,11]. Nevertheless, strict inclusion criteria often constrain these methods because they focus on a limited number of studies and specific intervention effects [12]. To address the limitations of traditional methods, there is a need for innovative approaches that analyze a broader scope of rapidly accumulating literature on PIs in NHs, integrating both macroscopic and microscopic perspectives.
Network analysis is an effective method for systematically exploring large datasets, identifying keyword relationships, and characterizing research trends [13,14]. By structuring knowledge, network analysis helps diagnose the current flow of research and suggests directions for future research [14,15]. Additionally, topic modeling enables the discovery of potential topics within textual data and the analysis of how these topics evolve [14,16]. Combining network analysis and topic modeling allows for a deeper understanding of research trends and helps identify knowledge gaps that conventional methods may overlook [14,16].
This study uses research abstracts as textual analysis data to explore the core keywords and topics associated with research trends in PIs within NHs. By identifying the knowledge structure and research gaps, the study aims to provide a comprehensive understanding of the current state of research and guide future studies to enhance the prevention and management of PIs in NHs.
Methods
Research design
This study aims to extract keywords and identify topic trends related to PI management in NHs using text network analysis and topic modeling (Fig. 1). Abstracts from relevant studies were analyzed to identify recurring patterns and thematic trends within this research field.
Overview of the research process. LDA, latent Dirichlet allocation.
Data collection
A systematic literature search was conducted for studies published from January 1970 to December 2023 using four electronic databases: PubMed, CINAHL, Embase, and Web of Science. The keywords “pressure injury” and “nursing home” were used. The initial search yielded 4,358 studies, of which 1,579 were duplicates. The remaining 2,779 studies were screened based on titles and abstracts, and those not meeting the eligibility criteria were excluded (n=922). The eligibility criteria excluded studies focusing on populations outside NHs (e.g., acute care hospitals or community settings), studies without PIs as the primary focus, grey literature, dissertations, reviews, and studies with incomplete data or unclear methods. A total of 1,857 studies were included in the final analysis. Two independent researchers screened the titles and abstracts and resolved disagreements through discussion to reach a consensus on study eligibility (Fig. 2). To ensure a clear distinction between NH and acute care hospital settings, studies were classified based on the research setting descriptors found in the title, abstract, or full-text descriptions. The included studies originated from diverse nations, such as the United States, Australia, Sweden, Spain, France, and Japan, reflecting a broad spectrum of NH environments and medical systems.
Process of data collection.
Data analysis
A total of 1,857 studies were analyzed using NetMiner version 4.5 (Cyram Inc.). Abstracts were uploaded in Excel format and processed for analysis. The analysis included data preprocessing, dictionary construction, keyword extraction, text network analysis, and topic modeling (Fig. 2).
Data preprocessing
During preprocessing, the abstracts were converted to lowercase, and the extracted terms were designated as nouns to capture the main concepts. Two researchers independently reviewed the extracted terms and developed a dictionary comprising defined words, a thesaurus, and stop words.
Top keywords and text network analysis
The researcher-constructed dictionary was implemented in the program to extract the top keywords from the abstracts, and their frequency, co-occurrence, and centrality metrics were analyzed to identify the most significant terms and relationships [14]. Keyword frequency highlights the most common terms across all documents, emphasizing their relevance to PI research. The top 30 keywords with the highest frequency were selected to represent key themes such as patient care, prevention, and wound management.
Co-occurrence analysis examines the relationships between keywords by measuring the frequency with which keyword pairs appear in the same document [17]. This analysis reveals thematic clusters and associations between frequently co-occurring terms, providing insights into the interconnectedness of key topics [14,17].
A text network analysis was conducted to further explore the structure and significance of the keywords within the network. A 2-mode document-word network was converted into a 1-mode keyword-keyword network, where connections represented the co-occurrence of keyword pairs [14,15]. Essential network characteristics, including density, average degree, and average distance, were calculated to understand the overall connectivity and spread of the keywords [13-15]. Centrality metrics were also analyzed to evaluate the importance of individual keywords in the network [18]. Degree centrality measures how frequently a keyword co-occurs with others and identifies core terms that represent the main topics of research [13,18]. Closeness centrality identifies keywords that are closely connected to others, facilitating the efficient spread of information and linking related terms [13,18]. Betweenness centrality highlights keywords that act as bridges between different clusters or subtopics, connecting distinct themes and ensuring the flow of information across networks [13,18].
This analysis combined frequency, co-occurrence, and centrality metrics to identify the most critical keywords and their roles in the research field [13-15]. These insights provide a comprehensive understanding of the thematic structure and interconnectedness of PI research, highlighting its core areas and broad implications.
Topic modeling
Latent Dirichlet allocation (LDA) was employed for topic modeling to identify key topics and associated keywords. This probabilistic approach estimates the distribution of topics within a corpus by analyzing recurring keyword patterns and clustering-related terms [19]. The parameters for the topic modeling process included α=0.01, β=0.01, and 1,000 iterations, ensuring robust and consistent topic extraction [16]. For example, if the keyword ‘assessment’ frequently co-occurs with terms like “risk,” “tools,” and “prevention,” LDA might categorize these into a topic related to preventive care.
Following LDA, multiple simulations were conducted to determine the optimal number of topics, guided by metrics such as perplexity and coherence scores [16,19]. Each topic was reviewed and categorized based on the most representative keywords extracted during the analysis. Appropriate names were assigned to reflect the thematic content of each topic.
To complement the topic modeling process, a temporal analysis was performed to assess the distribution and prevalence of topics across different periods. This analysis calculates the relative proportion of documents assigned to each topic within specific time intervals, providing insights into how the research focus has shifted over time [16]. Additionally, these trends were visualized using line graphs to highlight the changes in the prominence of topics, thereby enabling a clearer understanding of the evolving research landscape.
Results
Core keywords related to PIs in NHs
This study analyzed 30 keywords based on frequency, co-occurrence, and centrality metrics to assess their significance in research on NH PIs (Table 1). The most frequent keywords included patient (n=6,707), care (n=3,690), and wound (n=1,400), which emerged as central themes, underscoring the importance of patient care, wound management, and caregiving environments in PI research. These terms underscore the fundamental role of patient-centered approaches in managing PIs.
High ranked keywords related to pressure injuries in nursing homes
Co-occurrence analysis revealed that keywords such as patient (n=1,534), care (n=1,194), and facility (n=558) were frequently paired, illustrating the interconnectedness of caregiving and environmental factors in institutional settings. These relationships emphasize the importance of addressing both individual care and systemic factors in PI research.
Network analysis further demonstrated the centrality of the key terms within the research network. Keywords such as patient, care, and wound exhibited a high degree of centrality, indicating strong connections to other terms and their roles as core topics. The patient and care keywords had high closeness centrality, reflecting their pivotal role in facilitating information flow within the network. Additionally, these terms displayed high betweenness centrality, serving as bridges that connect various subtopics, ensuring cohesive relationships across thematic areas.
Integrating frequency, co-occurrence, and centrality metrics, patient, care, and wound emerged as pivotal keywords that shaped the structure and focus of NH PI research. These findings emphasize the interconnected themes of patient care, environmental factors, and wound management as core areas of study.
Topics related to PIs in NHs
Labeling of topics related to PIs in NHs
Topic modeling of research papers on PIs in NHs identified four distinct topics, categorized based on related keywords. These topics were labeled as Managing PI Complications (topic 1), focusing on care after the occurrence of PIs; Preventing PIs and Maintaining Skin Integrity (topic 2), which emphasizes strategies for preventing the development of PIs; Improving Organizational Strategies for PI Care (topic 3), addressing system-level support and management; and Managing Systemic Conditions in PI Care (topic 4), which explores individual patient characteristics contributing to PI risk (Table 2, Fig. 3).
Results of topic modeling
Topics related to PIs in nursing homes. PI, pressure injuries; MRSA, methicillin-resistant Staphylococcus aureus.
Trends in research on PIs in NHs
To examine trends in research on PIs in NHs, studies were categorized by decade, and the distribution of topics was analyzed based on the percentages provided.
In the 1970s, research was evenly distributed among topics 1 (33.33%), 2 (33.33%), and 4 (33.33%), with no studies focusing on topic 3 (0.00%). In the 1980s, the focus shifted significantly to topic 1 (56.25%), whereas topic 2 (12.50%), topic 3 (18.75%), and topic 4 (12.50%) had much lower proportions. In the 1990s, the distribution became more balanced, with topics 1 (18.91%), 2 (27.86%), and 3 (27.86%) maintaining similar proportions, while topic 4 (25.37%) also saw a noticeable share. By the 2000s, research on topic 3 (43.10%) had gained prominence, followed by topics 2 (25.10%), 4 (18.83%), and 1 (12.97%). In the 2010s, this trend shifted slightly, with topic 3 (34.83%) remaining dominant, followed by topics 1 (21.24%), 2 (23.30%), and 4 (20.63%). Finally, in the 2020s, the distribution became more balanced, with topics 1 (24.18%), 2 (24.18%), and 3 (33.13%) showing similar proportions, while topic 4 (18.51%) continued to have a smaller but consistent presence.
These findings highlight distinct trends of evolving research foci on PIs in NHs over the past few decades. Topic 1 peaked in the 1980s (56.25%) and decreased over subsequent decades, reflecting a declining focus on post-injury care. In contrast, topics 2 and 3 exhibited increasing trends. Topic 3 reached its highest proportion in the 2000s (43.10%), highlighting a growing emphasis on proactive and systematic PI management approaches. Topic 4 maintained a relatively stable proportion across the decades, ranging from 12.50% to 25.37%, suggesting consistent attention to individual patient characteristics and their impact on PIs.
Fig. 4 graphically represents these trends, illustrating a shift in research priorities over time. The results indicate an increased focus on preventive care and organizational processes, along with a more balanced distribution of topics in recent years.
Trends in research on nursing home pressure injuries (PIs) by period. Topic 1, Managing PI Complications; Topic 2, Preventing PIs and Maintaining Skin Integrity; Topic 3, Improving Organizational Strategies for PI Care; Topic 4, Managing Systemic Conditions in PI Care.
Discussion
This study provides a comprehensive analysis of research trends on PIs within NHs, focusing on core keywords and thematic areas through text network analysis and topic modeling. Topics identified in the analysis correspond to key areas of PI research: topic 1 represents post-injury management, topics 2 and 3 align with preventive care and organizational strategies, and topic 4 highlights patient risk factors. This categorization provides a structured framework for the subsequent discussion. These findings offer valuable insights into the evolving priorities of PI research, with an emphasis on the unique challenges and characteristics of NH settings.
The analysis revealed distinct trends over the decades, reflecting a shift in research focus. Studies in the 1970s and the 1980s predominantly addressed post-injury management, emphasizing strategies for managing PIs after their occurrence. This focus was likely driven by the limited resources and knowledge available at the time, necessitating reactive care approaches in NHs [20]. However, the decreasing emphasis on post-injury management in the subsequent decades suggests advancements in wound management techniques, improved clinical guidelines, and a growing emphasis on preventive measures [7,8,21].
In contrast, recent decades have witnessed a pronounced shift toward prevention and organizational strategies, particularly in the 2000s and the 2010s. Unlike acute care settings, NHs face unique challenges that make prevention critical [21]. NH residents often experience prolonged stays, reduced mobility, and chronic conditions, all of which increase their vulnerability to PIs [22]. These factors necessitate proactive measures including risk assessment tools, pressure redistribution devices, and staff education, which have been increasingly adopted in NHs [7,20]. These findings underscore the importance of tailoring preventive measures to the specific needs and constraints of NH environments.
NHs present distinctive environmental and systemic challenges that differentiate PI management from other healthcare settings. Extended care requirements for NH residents increase their exposure to pressure and shear forces, which are key factors in PI development [8,22]. This necessitates sustained preventive measures and individualized care plans. Staffing shortages and budgetary constraints in NHs can further impede the implementation of comprehensive PI prevention programs [20,22]. Addressing these systemic issues requires organizational strategies to optimize resource allocation. Additionally, NH residents commonly present with conditions such as dementia, malnutrition, incontinence, and reduced mobility, which significantly increase their PI risk [22,23]. Integrated approaches that combine patient-specific care and systemic interventions are essential to mitigate these multifactorial risks effectively [22,23].
The prominence of organizational strategies in the 2000s reflects growing recognition of these systemic challenges. Previous studies have shown that NHs with well-structured policies, interdisciplinary teamwork, and consistent staff training achieve better outcomes in PI prevention [6,24]. This study confirms that organizational-level interventions have become a dominant focus of recent research, further highlighting their critical roles in NH care.
Our findings have several practical implications for NHs. The increasing emphasis on prevention highlights the need for innovative strategies tailored to NHs, such as advanced pressure-monitoring technologies and early detection tools. These measures can help mitigate risks prior to the development of PIs. NHs must prioritize structured policies and staff education to address systemic challenges effectively [20]. Interdisciplinary collaboration and resource optimization are critical for improving the quality of care and reducing the incidence of PIs [4,24]. A stable focus on patient risk factors underscores the importance of individualized care. NHs must integrate personalized interventions into broader organizational strategies to address residents’ complex needs effectively [25].
This study has limitations owing to its reliance on abstracts, which may overlook nuanced findings in full-text articles. Future research incorporating full-text analyses could provide deeper insights. The selection and refinement of the literature may also have introduced bias, despite using a consistent strategy and expert input to ensure objectivity. Additionally, keyword categorization involved subjective judgment, although systematic reviews and expert validation minimized potential errors. Furthermore, even though the included studies originated from diverse nations, such as the United States, Australia, Sweden, Spain, France, and Japan, NH environments and medical systems vary significantly across countries, which may limit the generalizability of the findings. These results should therefore be interpreted with caution, as they may not fully represent the challenges or practices in lower-income or resource-limited settings. Future research should include a broader spectrum of nations and healthcare systems to enhance the applicability of PI management strategies globally. Finally, the study focused on publications up to a specific period, potentially missing recent advancements such as those involving artificial intelligence. Future studies should explore these developments to enhance our understanding of PI management.
Despite these limitations, this study offers valuable insights into the scientific knowledge structure surrounding PIs in NHs. By analyzing key keywords, identifying thematic topics, and examining trends over time, it provides a robust foundation for advancing nursing practice, education, and research, particularly in addressing the unique challenges faced by NHs.
In conclusion, this study underscores the evolving priorities of PI research, transitioning from reactive approaches, such as post-injury management, to proactive strategies, including prevention and organizational strategies. This explicit shift reflects the growing emphasis on systemic interventions and patient-centered care. These findings highlight the need for an integrated approach that combines patient-centered care with systemic and preventive measures. Future research should prioritize innovative methodologies and interdisciplinary collaborations to further enhance PI management in NHs, and ultimately improve care outcomes for this vulnerable population.
Notes
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (2021R1G1A1091862) (2022R1A2C1004542).
Ye-Na Lee is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.