2025/08/11 by Josef Trapani, Adrienne Grech, Anthony Scerri · 1 voice · 1 citation
Health Professions · Medicine · #Family and Patient Care in Intensive Care Units #Geriatric Care and Nursing Homes #Respiratory Support and Mechanisms
paper · pdf · doi:10.1111/nicc.70127
openalex publication_date 2025/08/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/14
Critical care research plays a vital role in advancing clinical practice, improving critically ill patients' and their families' outcomes, and enhancing healthcare workforce well-being. Nursing in Critical Care (NICC) serves as an essential platform for disseminating high-quality research that informs evidence-based critical care practice. As we approach the journal's 30th anniversary, it is timely to take a look back to celebrate the research published in the journal over these years and another look forward to identify possible gaps and areas for future research. This editorial presents the key points of a research trend analysis conducted on original research articles published in NICC since 2003,1 in an effort to provide valuable insights into the most prevalent research themes, ongoing challenges and gaps in critical care research published in the journal. The analysis was supported by natural language processing (NLP), machine learning techniques and large language models to identify and categorise research topics systematically. By examining research frequency, distribution and evolution over time, this analysis highlights research trends, gaps and possible future priorities in critical care nursing. All articles that were categorised under ‘research article’ ‘service evaluation’, ‘short research article’ ‘quality improvement’, ‘practice development’ and ‘case study’ from 2003 to the first quarter of 2025 were manually extracted and compiled into a single document. A Python-based NLP script (presented in the Data S1) was developed, with the aid of ChatGPT-4o, to process, analyse and visualise research trends. The analysis followed a structured approach, beginning with data extraction and pre-processing, followed by topic identification using machine learning and concluding with trend visualisation. To refine the studies and ensure that only primary studies were processed, the script included an automatic exclusion of review articles, commentaries and editorials. The first step involved extracting and pre-processing data. Key elements, such as the title, abstract, keywords and publication year, were retrieved from each study. The extracted text was cleaned by removing common English stop words (e.g., ‘the’ and ‘and’) as well as domain-specific stop words (e.g., ‘nurse’ and ‘ICU’) that might skew the analysis. Additionally, lemmatisation, a process that standardises words to their root form (e.g., ‘nurses’ → ‘nurse’), was applied to ensure consistency. The second step focussed on identifying research topics using BERTopic, an unsupervised machine learning model specialised in text clustering [1]. BERTopic was used to group studies into clusters based on their textual similarity. It then extracted key terms that represented each research topic and ranked topics by frequency. To capture meaningful word combinations, bigrams and trigrams were generated, allowing the analysis to detect commonly co-occurring phrases. These word combinations (bigrams and trigrams) were extracted from the context of each research study and were based on their frequency and distinctiveness and not on syntax or semantic completeness. This approach enabled the identification of the most frequently investigated topics, providing an indicative measure of the research published in the journal. The next step involved analysing and visualising the most frequently studied research topics. The top 20 research topics by frequency were identified and ranked based on their occurrence across studies (Columns A, B and C in Table 1 below). Each identified topic was then interpreted and categorised into broader research domains, such as critical care management, healthcare workforce well-being, and infection prevention and control (Column E in Table 1). These top 20 research topics were then grouped into overarching themes (Column F in Table 1) with the help of Google Gemini, which also determined the frequency and percentage in each overarching theme (Table 2) and helped to identify research gaps. Patient management & clinical outcomes * Ventilator weaning and respiratory care * Sleep and rest in critical care * Pain assessment tools * Delirium management and cognitive burden * Early mobilisation in ICU * Airway and respiratory equipment use * Post-extubation complications * ICU recovery and long-term outcomes * Cognitive dysfunction and delirium Family-centred care * Family-centred care * Neonatal family-centred care Education & training * Education and practice development * Simulation-based education Staff well-being & workload * Staff well-being and moral distress * Nursing workload and staffing * Staff mental health and well-being Safety & harm prevention * Patient safety and skin integrity * Infection prevention in mechanical ventilation * Ethical issues in ICU safety End-of-life care * End-of-life decision making The compiled single document consisting of the 737 retrieved articles was translated into a Microsoft Excel spreadsheet for further analysis related to the specific population distributions. The spreadsheet was analysed using Claude Opus 4 NLP techniques and thematic categorisation of paediatric, adult and both populations in critical care. A textual classification system based on MeSH terms and synonyms related to these populations, such as ‘paediatric’, ‘neonate’, ‘children’, ‘infant’, ‘adult’ and ‘elderly’, was developed. A total of 54 and 47 terms were identified for paediatrics and adults, respectively. If no keywords for paediatric, adult or both populations were present, articles were grouped under ‘unspecified’. This phase of textual analysis consisted of (i) text pre-processing by removing common stop words, filtering words with less than four characters and excluding methodology-related terms; (ii) term extraction by calculating the frequencies for each population group and population-specific themes; and (iii) identification of population-specific themes. A total of 810 articles were published in NICC under the categories of ‘research article’ ‘service evaluation’, ‘short research article’ ‘quality improvement’, ‘practice development’ or ‘case study’ between 2003 and March 2025. Seventy-three articles were excluded as they were automatically categorised as reviews, commentaries or editorials, leaving 737 primary research studies for analysis. Figure 1 displays the percentage of papers within these criteria that were published during each year indicating that the percentage has increased exponentially over the past 5 years2. This research trend analysis considers the top 20 most frequently studied topics, as identified through NLP, in the primary research articles published in NICC. These constitute a total of 368 primary research articles, with the remaining 369 primary research papers tackling topics that did not fall under any of these 20 most frequent ones. Table 1 ranks these 20 most frequent topics, provides a possible interpretation of each generated topic and collapses the research topics into overarching themes. Table 2 presents these overarching themes and topics within each theme, together with the overall frequency of that theme. As shown in Table 2, almost half of the papers related to the top 20 topics were dedicated to Patient Management & Clinical Outcomes; this is unsurprising considering that holistic nursing care in critical care settings is a main focus of the journal. Prominent in this overarching themes are areas of care to which the journal included special issues, such as delirium assessment and management (Volume 26 Issue 3); respiratory care and ventilation weaning (Volume 29, Issue 6); the psychological impact of the ICU environment (Volume 25 Issue 2); early mobilisation in ICU (Volume 29 Issue 3); as well as survivorship (Volume 26 Issue 2) and rehabilitation (Volume 28 Issue 4) during and after critical care. This overarching theme also includes topics related to patient comfort, such as sleep quality and pain assessment and management in critical care settings. Family-Centred Care in both adult and neonatal ICUs is the second most popular category by frequency; it is, therefore, timely that the journal currently has an open call for papers for a special issue in this area following another family-focussed issue a few years ago (Volume 27 Issue 3). Nursing in Critical Care's target audience is not only critical care clinicians and researchers, but also educators and managers. Consequently, the third and fourth research themes by frequency focus on Education & Training (including practice development and simulation-based education) and Staff Workload and Well-being (including staffing issues and moral distress). These reflect increasing attention given to skills development and safeguarding critical care professionals' mental health and efforts to prevent, detect and address burnout, respectively. Of note, NICC published a special issue on leadership and management in intensive care (Volume 28 Issue 2) and another one which partly focussed on well-being and moral injury (Volume 29, Issue 6). In 2021, we also published an issue with a special focus on workforce considerations in intensive care (Volume 26, Issue 6). Congruent with the notion of failure to rescue [2], 10% of this body of literature focussed on some aspect of Patient Safety & Harm Prevention, including skin integrity, infection prevention and the use of patient restraints, as well as a special issue on quality and safety (Volume 20 Issue 4). In 2019 and 2020, NICC published special issues on bereavement care (Volume 24 Issue 4) and on dying, death and end-of-life care (Volume 25 Issue 5). Correspondingly, papers on End-of-Life Care and decision-making around it contributed to around 6% of the total original research articles published in NICC during the period under exploration. Most of the papers published in NICC were reports of original research studies, followed by various types of literature reviews and a much smaller proportion of case studies, quality improvement and editorial/commentary papers. Among the research papers, 30% adopted a qualitative approach, and 58% adopted a quantitative approach, while the NLP analysis could not determine the research design of the remaining 12%. Among quantitative papers (n = 433), half were cross-sectional studies (which is reflected through the various survey studies published in the journal), 27% were descriptive, 12% were randomised controlled studies, and the remaining 13% were cohort or case–control studies (Figure 2). Textual analysis conducted by the large language model Claude Opus 4 revealed that 58.6% of the papers did not specify a population group. Papers that specifically indicated that they focussed on adult critical care comprise 26.7% of all studies, significantly outnumbering paediatric3 papers (10.2%) by a ratio of 2.6:1. The ratio of adult to paediatric papers is probably much more highly skewed in favour of adult critical care because most papers that do not specifically indicate an age group actually focus on adult critically ill patients, their families or those providing them a service. Of note, only 4.5% of articles specifically mentioned addressing both paediatric and adult populations. Although it is not surprising that research focussing on adult critical care is predominant, it is important to note that the proportion of paediatric and neonatal-focussed papers has substantially increased over the years, as shown graphically in Figures 3 and 4. Indeed, if one were to focus on the last 5 years and exclude articles with an ‘unspecified’ population, the proportion of paediatric papers increases to 26% (Figure 3). This reflects the efforts of the current editorial team to promote critical care across the lifespan, through the recruitment of peer reviewers with expertise in neonatal and paediatric critical care, as well as adding more persons with such expertise to the journal's editorial advisory board. From among the paediatric-focussed papers, there is a considerable proportion (41%) focussing on neonates aged between 0 and 28 days, whereas the much larger 4–12-year-old age group is the focus of 57% of the papers (Figure 5a). Similarly, in adult-focussed papers, 42% specifically address older adults (65 years or older) (Figure 5b), which again was made possible with the engagement of reviewers with expertise in the care of older adults and reflects the journal's celebration of critical care across the lifespan. Thus, neonatal care dominates paediatric research (41%), while the care of older adults is prominent in adult research (42%). Conversely, the least represented age groups are young and middle-aged adults and adolescents, which is not surprising since relatively fewer persons from these age groups require admission to intensive care, and their length of stay tends to be shorter [3, 4]. A closer look at the focus of the research among adult and paediatric patients reveals some similar and some different priorities. Approximately 41% of paediatric papers specifically focus on neonatal care, highlighting the critical nature of early life interventions, whereas sedation and delirium assessment and management, and mechanical ventilation are predominant in adult research, reflecting complex intervention needs. Education and practice development are the predominant themes in both populations; this is possibly because the abstracts of studies about both population groups typically include recommendations for education or practice development. Family involvement is given importance in both adult and paediatric papers, but a family-centred approach is significantly more predominant in paediatric articles (45.3% vs. 17.3%). In paediatric articles, there is an emphasis on end-of-life decision-making in critically ill neonates and children, while adult research focuses more on delirium management. Table 3 summarises the most popular themes in adult and paediatric papers published in the journal. Although this analysis is meant to provide a schematic overview of the trends and topics in NICC publications, rather than an exhaustive analysis, some areas appear to be either missing or only briefly addressed. These are briefly discussed in this section. Although topics like ‘airway and respiratory equipment use’ and ‘simulation-based education’ are prominent, there is less explicit focus on broader technological advancements, such as the application of artificial intelligence (AI) in critical care nursing, advanced monitoring systems or the implementation and evaluation of new digital health tools. Nonetheless, in recent years the journal did publish a research study on telemedicine [5] and a guest editorial on ICU dashboards [6]. A special issue related to this area is also being considered. The collection heavily features respiratory care and delirium management. However, research specific to nurses' role in other prevalent critical care conditions, such as sepsis, cardiovascular emergencies, neurological critical care (apart from delirium), renal failure management or severe trauma, appears to be much less prominent. Nonetheless, the journal regularly publishes papers related to cardiac intensive care nursing (including a special issue in Volume 27 Issue 2) and occasionally case studies on the care of patients with specific conditions [7-9], including relatively rare ones [10-12]. While staff well-being and workload constitute an important category in this collection, no specific categories focussing on the dynamics of interdisciplinary teamwork, communication strategies among members of different healthcare professions or their direct impact on patient outcomes or care efficiency feature among the 20 most frequent research topics. This constitutes another area that may be addressed in future submissions. The list of most popular topics includes subjects like ‘Pain Assessment Tools’ and ‘Delirium Management’, which indirectly touch upon medication. However, a broader category dedicated to medication administration safety, complex polypharmacy management, drug interaction monitoring or the nursing implications of specific pharmacologic interventions is not prominent. Exceptions to this include a small number of papers on medication administration [13-15]; studies on ICU nurses' perspectives on the co-administration of multiple intravenous medicines [16] and medication error [17]; and the British Association of Critical Care Nurses (BACCN) position statement on non-medical prescribing [18]. Although the journal has a consistent track record of publishing research from diverse settings and countries, papers addressing health disparities or focussing on equitable access to critical care services do not feature prominently. A notable exception is a critical commentary on the challenge for low-middle income countries during the COVID-19 pandemic [19], which should hopefully encourage other authors to publish empirical, review or commentary papers on this important topic. Also does not feature among the most frequently tackled topics, reflecting a general delayed attention to this imminent and vital issue until relatively recently. A positive development against this trend is the journal's publication of an empirical study [20], a guest editorial [21] and an entire special issue dedicated to environmentally sustainable critical care earlier this year (Volume 30 Issue 1), which coincided with the formal launch of the Intensive Care Environmental Sustainability Recipe Book [22]. This editorial celebrates the 30 years since the first issue of NICC by providing a brief research trend analysis which hopefully provides valuable insights into the evolving landscape of critical care nursing research, helping to identify the most popular research topics, research gaps and priorities for future research. In so doing, it provides an example of the use of NLP and large language models in facilitating the identification of trends in large sets of documents and the importance of acknowledging their use in academic outputs. It is believed that this encourages authors who have not made use of such tools to consider doing so transparently, and those who have, to acknowledge the manner in which they were used to enhance research integrity. Although it is not meant to be an exhaustive analysis, this exploration has several strengths, through which it provides a good indication of the publication trends in the journal. First, the evaluation and analysis of trends employed a combination of large language models that complemented each other in analysing the same data set. Second, the analysis was performed on an extensive summary of all the original articles published in NICC over the past two and a half decades. Third, three researchers reviewed the outputs of the AI models for consistency, clinical relevance and contextual interpretation. Limitations of the analysis by population group include the dependency of the textual analysis on keywords that explicitly mention population; therefore, some studies which implicitly refer to specific populations but did not use certain keywords were classified as ‘population undetermined’ and thus excluded from the final percentages. Also, this part of the analysis on the interpretation of a large language model Opus 4) of all Although the analysis does into the different age groups of critically ill adults, and an exhaustive analysis of the between paediatric and neonatal critical care research published in the journal was the of this we acknowledge the between these two this may be the focus of a similar analysis. the efforts to focus on primary research studies, removing studies through NLP, the could have automatically primary studies that included the term in the However, a of the indicated that such were the analysis of the gaps in the literature was based on the 20 most frequently some of these gaps may indicate that they did not feature in the list rather than being important of such an analysis is to that this less frequently critical care such as the application of advanced and of digital interdisciplinary health and Additionally, these insights may help to and practice in critical care nursing, evidence-based decision-making to patient care and healthcare workforce well-being, which in the most popular research themes and topics. trends in research also interdisciplinary allowing researchers, clinicians and to together in addressing complex healthcare such as staff burnout, infection prevention and control and family-centred care. As the of critical care nursing to future research to that patient workforce challenges and healthcare system at the Nursing in Critical Care of publishing high-quality papers related to critical care practice, research, education and management, and ongoing trend analysis an essential for advancing practice and in this that the research priorities with the challenges in critical care settings. The authors no of Data not to this as no were generated or during the current study. Data The is not for the or of any by the than missing should be to the for the