The current research involved the use of several analytical tools to support data processing, bibliometric analysis, visualization, and retrieval of altimetric data. Microsoft Excel, developed by Microsoft, was used for preliminary data cleaning, removal of duplicates using DOI matching, descriptive statistical analysis and preparation of datasets for further processing. Advanced bibliometric analysis and scientific mapping were performed with Biblioshiny, a web-based graphical interface built in Bibliometrix package in R programming environment. Biblioshiny facilitates full bibliometric analysis, including performance analysis, co-occurrence network construction, thematic evolution and visualization of research trends of large datasets. Altmetric Attention Scores were collected using automated data extraction techniques from the App Script functionality integrated into the Google Sheets, cloud-based platform, offered by Google. The process required for scripting enabled DOI-based retrieval and organisation of altmetric data to ensure consistency and reproducibility of the dataset.

Altmetric indicators were extracted using the DOI through Google Sheets App Script functionality integrated in the cloud environment provided in Google. The Altmetrics database was queried by the script using unique DOI identifiers and Altmetric Attention Scores and associated engagement indicators were extracted. This extraction process minimized manual error and ensured consistency in altmetric data integration to enhance transparency and reproducibility, the App Script workflow developed for DOI-based Altmetric data retrieval has been made publicly available through a GitHub repository ( https://github.com/sangeeta25496-alt/MaternalMalnutrition-AltmetricWorkflow).

It was observed that the articles were authored by either a single author or multiple authors. To study the authorship pattern in the concerned set of literature, the publications were divided into two groups: single-author publications, publications with a single author and multiple author publications, and publications with two or more authors. The year-wise distribution of the number of publications, citations, and average citations per publication across the year is presented in Table 1.

The authors tried to normalize the citations across publications published across the studied period. A suitable year can be defined as the number of years a publication is available for citations. For instance, the citable year for an article published in 2015 is 10 years, for an article published in 2016 is nine years, and for an article published in 2024 is one year, as we have taken the data in 2024. The Total citable year is defined as the total number of available citable years for all articles published in a given year; this is calculated by multiplying the number of publications by the citable year. Normalised citation is a measure of how efficiently the publication of a given year accrues citations in proportion to its outflow time. There is a normalization done, which helps mitigate the bias of older publications getting time to accrue more citations than newer ones. The formula used to calculate the Normalised citation is given in Equation 1. Normalised Citation = Citations recived in a particular year ( Citable year for a particular year ) ∗ ( Number of publications in that year ) (Equation 1)

Table 2 represents the Normalised citations received across each studied period. It can be observed that the Normalised citation score is highest for 2015, 2016, and 2017, of which 2016 topped the list, which shows that articles published during 2016 are more impactful than articles published during any other period.

RGR and DT provide an understanding of the growth rate and dynamicity of publications. Higher RGR values tend to increase the growth rate of publications, whereas lower DT values indicate rapid growth in publications. The doubling time (DT) is defined as the time required to double the total number of articles. The present study used the formulas mentioned in Equations 2 and 3 to calculate the RGR and DT. ^{37– 39} In the calculation of the RGR, W1 is the natural logarithm of the initial number of papers, W2 is the natural logarithm of the final number of papers, T1 is the initial year, and T2 is the final year. DT is inversely proportional to the RGR. In Equation 3, 0.693 is the natural logarithm of the two variables. RGR = ( W 2 − W 1 ) ( T 2 − T 1 ) (Equation 2) DT = 0.693 RGR (Equation 3)

Table 3 demonstrates the year-wise trends in the Relative Growth Rate (RGR) and Doubling Time (DT) for publications between 2015 and 2024. The total number of publications has grown significantly from 344 in 2015 to 4,740 in 2024. During the emergent phase (2016–2018), RGR was high (0.716 in 2016) and declined to 0.338 by 2018, whereas DT increased from 0.968 to 2.051 years. This indicates the rapid growth that is typical of a nascent research field. In the transition phase (2019–2021), RGR steadily decreased from 0.267 to 0.194 and DT increased from 2.600 to 3.581 years, reflecting slower growth as the field began to stabilize. The mature phase (2022–2024) showed continued declines in RGR (0.165 in 2022 to 0.121 in 2024) and steeper increases in DT (4.205–5.741 years), suggesting stabilization as publication growth slowed. These trends highlight the progression of the domain from a rapid initial growth to a stable and consolidated stage, following the typical lifecycle of a research field.

The global distribution of literature production is depicted in Figure 3, with the intensity of deep blue indicating higher publication output from specific regions. Regional contributions highlight diverse scientific engagement across the globe. In Latin America and the Caribbean, Brazil has contributed the most, followed by Mexico. Ethiopia and South Africa are the primary contributors to Sub-Saharan Africa, while India and Pakistan dominate South Asia. Iran and Saudi Arabia represent the Middle East and North Africa, respectively, and the USA and Canada lead in North America. In East Asia and the Pacific, China tops the list, followed by Australia, while the UK and Spain are the major contributors in Europe and Central Asia. Globally, the USA has emerged as the top contributor, followed by China, India, Ethiopia, and the UK, underscoring the geographic breadth of research in the domain.

The corresponding authors, typically the principal investigators or senior researchers, ensure the accuracy and integrity of a study. Publications are categorized as single-country publications (SCP), involving authors from a single country, or multiple-country publications (MCP), where authors collaborate across nations. Figure 4 illustrates the distribution of the corresponding authors by country, along with the MCP and SCP contributions. The USA leads the corresponding authorship, followed by Ethiopia, India, and the UK. Notably, MCP accounts for over 50% of the publications with corresponding USA-based authors, indicating a high level of international collaboration. In contrast, Ethiopia and India show MCP percentages below 20%, reflecting a stronger focus on local collaboration. Countries such as the UK, Australia, and Bangladesh also exhibit high MCP proportions, surpassing 50%, suggesting a significant emphasis on global research partnerships.

Table 4 shows the year wise distribution of Highly Cited (HC) publications. The publications were classified into five groups based on citation accrual as more than 1000 citations, 500-1000 citations, 100-500 citations, 1-100 citations, and uncited publications. It was found that publications with extremely high citation levels (>1000 citations) were very focused in earlier years of the study period especially between 2015 and 2017, which shows the cumulative citation advantage of the longer exposure to publication. Likewise, articles in the 500-1000 and 100-500 citation categories were found most in the first years, which showed citation maturity effects and time-dependent aspect of academic impact. Most publications in all the years are in the 1-100 range of citation, which shows the active academic involvement and average exposure to the research. Conversely, the number of uncited publications is significantly increasing since 2022, especially in 2023 and 2024, which is in part due to the recency of the publications and lack of time to cite new publications.

It was noticed that ten publications had received more than 1000 citations. Table 5 represents most HC publications in the data set. The top ten HC publications were published in 2015-2020 with citations of 1,511 to 6,153. The majority of the ten top works were published between 2015 and 2017 and it suggests that older publications were more likely to receive citation accrual because of a more extended period to cite. It was observed that all HC publications were published on The Lancet journal, hence it demonstrates the dominance of this journal in disseminating influential research in this area. The ten most cited papers in the dataset of present study have come from the Global Burden of Disease (GBD) study, which were published between 2015 and 2020 with citation counts ranging from 1,511 to 6,153. ^{40– 49} They provided a systematic way to understand global health trends over time and across different countries. A major theme across these publications was the burden of maternal and child malnutrition. The GBD 2013 risk factor study founded that child and maternal malnutrition was one of the six leading risk factors worldwide. It caused 1.7 million deaths and 176.9 million DALYs in 2013. The highest burden was in sub-Saharan Africa and South Asia. ⁴⁴ Later studies confirmed this finding. The GBD 2019 risk factor analysis ranked child and maternal malnutrition as the leading risk factor for global DALYs. It accounted for 295 million DALYs, or 11.6% of all DALYs worldwide. ⁴⁵ Iron-deficiency anaemia was the leading impairment by number of individuals affected globally, reaching 2.36 billion people in 2015. ⁴¹ Among children under 10 years old, the three most important risk factors were all related to malnutrition. This included childhood wasting, stunting, and underweight. ^{42, 46} These papers also showed how maternal health connects to child survival. Maternal disorders remain leading causes of death and disability in women of reproductive age. Nutritional deficiencies in mothers directly contributed to poor outcomes in newborns. ^{41– 47} The GBD 2015 mortality study noted that neonatal preterm birth complications and maternal disorders were among the causes showing significant declines, yet they still account for substantial burden in low-income regions. ⁴⁸ The GBD 2015 non-fatal analysis highlighted that in low SDI countries, nutritional deficiencies accounted for a large portion of disability. Iron-deficiency anaemia alone affected 2.36 billion people globally in 2015. ⁴⁹ The study by Foreman and colleagues projected that child malnutrition risks would continue to drive health disparities in sub-Saharan Africa through 2040, unless addressed through deliberate policy action. ⁴³

Despite overall improvements in global health, malnutrition persists. This highlighted the need for continued investment in maternal and child health. Key interventions include better antenatal care, micronutrient supplements, and programs that improve food security and early childhood nutrition. ^{42– 48}

The topic dendrogram in Figure 5 displays the hierarchical relationships among the various research themes. A dendrogram was generated by clustering the authors’ keywords using Multiple Correspondence Analysis. Concepts are connected through a hierarchical structure, forming a cluster according to the similarity, and connecting with another cluster through the axis. The shorter axis shows a narrower relationship, and the larger branches show a broader relationship. The vertical axis shows the similarities and dissimilarities among topics. The larger vertical axis shows the distance relationship between concepts, and vice versa. In this case, the focus lies on nutrition, malnutrition, maternal-child health , and related public health themes. The vertical axis of the dendrogram lists the key research terms, whereas the horizontal axis quantifies the degree of similarity or dissimilarity between these terms. Terms that merge at lower heights exhibit close conceptual relationships, whereas those connected at greater distances signify broader, less direct associations.

The four-quadrant approach of the thematic map in biblioshiny provides the idea of thematic areas of research in the concerned domain. Four thematic areas were embedded in the two-dimensional Figure as shown in Figure 6. The Horizontal axis shows the density, which indicates the level of development of the thematic areas compared to the internal association of keywords. The second aspect is centrality, which measures the importance of the thematic area by means of an external association among the keywords. ⁵⁰ Centrality represents the volume of citations associated with the theme, and density represents the number of publications associated with theme. ³⁵ The upper-right side of the quadrant is the motor theme, which represents the central and developed areas of the domain. The Upper left quadrant is the Niche Theme, which represents very specialized areas of the domain; the lower left quadrant of the Figure represents the Emerging or Declining theme, which represents the emerging or declining areas of the domain; and the lower right quadrant represents the Basic Theme, which represents the underlying areas of research that are regarded as the base for research development. ^{51, 52}

A thematic map is shown in Figure 6, which was created using the keywords of the authors as inputs. The map was generated with specific parameters: a minimum cluster frequency of 5 per thousand, 250 words, 5 labels, and a label size of 0.35. Its thematic map is an informative and insightful visualization based on the authors’ keywords, showing which key themes they refer to and how they are related. The thematic map provided a comprehensive analysis of research trends and themes in the domain of maternal nutrition. This map, structured into four quadrants—Motor Themes, Niche Themes, Emerging or Declining Themes, and Basic Themes—illustrates the density and centrality of various thematic areas, based on scientometric data. Below is a detailed interpretation of each quadrant and the identified clusters.

The growing importance of social media has ignited scholarly communication in several ways. In this regard, alternative metrics or altmetrics provide a better understanding of public discourse on a particular topic. Traditional citation-based metrics limited their scope to measure the academic impact of scholarly output, but the alternative metrics empower us to capture and understand the social impact of scholarly output. Altmetrics measures scholarly output on various social media platforms. The altmetrics attention score (AAS) on various platforms gives us the idea of how a particular topic is discussed among the public, policymakers, and other stakeholders. In other words, altmetrics empowers us to study the real-world impact of scholarly output in real time scenario. ^{53, 54}

The study used altmetrics data from altmetrics.com through an API call on Google Sheets. The overall AAS is the weighted average of the article’s engagement on various social media platforms, calculated using an automated algorithm. ⁵⁵ It can be observed from Figure 7 that, out of all platforms, the maximum weightage has been given to “News” (8) followed by “Blog” (5). Online platforms such as “Policy document”, “Clinical Guidelines”, “Patent, and “Wikipedia” have a weighting of three.

Out of 4741 articles, 4589 articles have a digital object identifier (DOI). The DOI of 4589 articles was processed in Google Sheet, and the overall AAS and the individual attention score of each article across various platforms were retrieved through suitable API calls on Google Sheet. It was found that out of 4589 articles, 2529 had zero overall altmetrics scores. It was also found that a total of 763 such articles have zero overall attention score but have at least one attention score in any one of the platforms, since the overall AAS is the weighted average of the attention received across various platforms. The year-wise individual AAS across each platform of each of the articles across all platforms have been summed up. Table 6 shows the year-wise attention scores of all the articles across various platforms. The Data of Table 6 reveal an evolving pattern of individual AAS across the various platforms. A surge in Mendeley readership can be noticed during the years 2015–2016, but a steady decline in AAS over Mendeley can be noticed from the year 2017 to the recent year, which suggests that the particular topic has gained saturated attention or a shift in academic focus during this period. A significantly high score on Blogs and Twitter suggests continuing public discourse on the topic of concern. However, the AAS on Blogs and Twitter can be noticed during the year 2018, 2021 and 2024, which suggests that research conducted during this period gathers more public interest. The AAS on News platform was highest during 2018 and then showed a declining trend, which suggests that research during the year 2018 had potential mainstream attention. The score on policy documents was high during the initial period, which suggests that research conducted during the earlier period was considered for policy discussions, and the gradual declining trend on social media platforms such as Facebook suggests that research during the earlier period attracts more informal attention to social media sites like Facebook. An immediate spike in Wikipedia mentions can be noticed in 2016, but a declining trend can be noticed afterward, which indicates reduced public references or edits related to the topic. A high score on patents can be noticed during the early period, but a declining trend can be noticed afterward, which suggests that research during the earlier period was used to create innovation or technological development in the concerned field.

Despite of the extensive analysis and interpretation, the present study also has some limitations. The first limitation of the study is, the study solely depended on SCOPUS database for primary point of data collection, which may potentially ignore the publications indexed in other major databases. Secondly the study is limited for the period 2015 to 2024, which may potentially ignore the publications published beyond this period.

Third, altmetric analysis quantifies the impact of research by the attention it receives in online and social media and thus depends on technological and dissemination issues. The AAS cannot be regarded as the only measure of academic impact. The social media platforms regularly update their algorithm for better content visibility to get the competitive advantage, which directly influence the AAS. Moreover, fluctuations in the popularity of the platform and alterations in the Altmetric policies of data collection could affect the trends in attention in the years of publication. Thus, Altmetric indicators must be viewed as an indicator of public and media attention and not as the direct indicator of scientific quality and the long-term academic impact.