Environmental tobacco smoke and children’s health: a bibliometric and altmetric analysis of 100 most cited articles

Analysis of the top 100 cited articles on exposure to ETS and its impact on children’s health provides a varied yet persuasive read. This study links conventional indicators of bibliometrics with the modern digital dissemination measures for studies relating to ETS and CH. Currently, they appear to have discrete but reciprocal parts in assessing the broadcasting and influence of these publications.

One of the most striking features of the list is papers that appeared in journals with a low IF garnered substantial citations, whereas articles emerging in high-IF journals received limited references. The Pediatrics journal had the maximum number of articles (n = 12, JIF 9.703), whilst the British Medical Journal with maximal IF 93.333, presented only four studies. This suggests that citations are more dependent on the content and scientific ‘popularity’ of the research topic among researchers than the JIF. This study observed 33 articles with 100 or greater citation counts, thus making them citation classics [16]. They were cited between 100 and 641 times when the evaluation was employed with Scopus. A comparison across multiple data sources revealed variations in citation numbers; citations varied between ranges of 41–641 (Scopus), 35–549 (WoS), and 38–984 (Google Scholar). This difference underscores the purport of selecting a relevant scientometric database. Scopus provides a wide breadth of journals (n = 12,850) than WoS (n = 8,700) and quicker citation analysis. WoS and Google Scholar were not used as benchmark data sources for numerous grounds. In WoS, missing and incorrect references are major issues. Google Scholar includes citation data from books, preprints, theses, and dissertations which may influence the evaluation of the top publications [17]. Interestingly, two highly cited papers by Weitzman M on “Maternal Smoking and Childhood Asthma” and “Maternal Smoking and Behavior Problems of Children” were only found in Scopus and Google Scholar but not in WoS. It is worth noting that while citation counts do not delineate the study quality, it imitates its acclaim within the research community and impact on shaping discussions, controversies, practice guidelines or further investigations [18].

Although older literature is likely to be more frequently referenced, we observed a significant inclination towards recently published articles, with 33 papers that were released within the last decade. This can be attributed in part to the increasingly prominent role of digital platforms in evidence-based medicine, enabling manuscripts to explore novel concepts and guide future research trajectories. Interestingly, over the years the number of co-authors has risen substantially, with a preponderance of publications having more than three authors. A possible explanation could be increased awareness and interest among researchers of numerous institutes and countries about the potential benefits that studies in the purview of ETS could provide in children’s health. The average number of researchers per publication was 6.19. Frank D. Gilliland, a leading investigator in air pollution research, respiratory health, and gene-environment interactions, was on top of the list with six articles and a mean citation density of 29.31. In this analysis, it was observed that authors tend to collaborate quite frequently with authors affiliated with the same university or country. Frank D. Gilliland and Kiros Berhane had maximum collaborations with researchers. More coalition amongst investigators can be expected in the future.

As evidenced by the present study and in concordance with other bibliometric studies in varied fields, the majority of studies stemmed from academic institutes in the United States. Countries with a stronger economic background are inclined towards biomedical research, perhaps due to better scientific resources and funding. Despite the high prevalence and fatalities associated with exposure to ETS among children in low- and middle-income nations, there were limited population-based investigations performed within these regions. This study recognized a trend towards collation between the United States and several other nations, including the United Kingdom, Sweden, Italy, Netherlands, Poland, Germany, Switzerland, Greece, Denmark, Spain, China, Canada, Australia. Notably, among the top 100 cited articles, there were only two randomized controlled trials and five systematic reviews, while narrative reviews dominated with a count of 32. It is important to acknowledge the challenges of conducting randomised controlled trials for hazardous exposures like ETS even when trying to implement beneficial interventions. Furthermore, with the colossal size of publications, researchers may incline to consolidate and synthesize the existing information on a topic in the form of a literature review. Though Cochrane reviews have been internationally acclaimed as the highest level of the evidence base, they could ensure only one position in this study. A plausible explication of the lower citation counts could be that they are yet to attain a substantial age of publication. Fifty-five percent of the research papers were observational (cohort or cross-sectional). This finding could be attributed to the relatively lower resource requirements and costs associated with these study designs. Fundamental explorations in the etiopathogenesis of ETS have emanated from this study design. As the evidence-based philosophy is being propagated globally, it is essential to prioritize meticulously planned high-quality clinical studies on ETS and CH. Urgent attention must be directed towards conducting large longitudinal studies that span from preconception until childhood to gain a better understanding of how exposure to ETS impacts subclinical childhood health outcomes, such as neuropsychologic impairments. Additionally, large-scale case-control studies are required to investigate gene-environment interactions for relatively uncommon diseases like childhood malignancies. However, there are two challenges present within this field: exposure misclassification and statistical methodologies required for dealing with intricate interactions comprising multiple dimensions. Future research efforts could immensely benefit from using archives of exposure biomarkers which hold crucial information on prenatal and childhood determinants of adult diseases. While the primary target organ for ETS exposure is the lungs it comes as no surprise that a considerable number of studies (n = 31) focussed on respiratory outcomes such as asthma, wheezing, pneumonia, acute respiratory infections, and lower respiratory infections. There exists a substantial amount of evidence to support the causal relationship between exposure to ETS and respiratory ailments as compared to other conditions. There was a scarcity of studies assessing the association between ETS and atopic eczema (n = 9) or otitis media (n = 8). Similarly, the number of articles about ETS and snoring, and obesity were also low. The relationship between ETS and childhood dental caries is an area of research that is expanding. Furthermore, the expeditious growth of DNA methylation has aided the ranking of epigenetic papers, a part of Precision Environmental Health, to gain notable traction in the past ten years. It is paramount to take cognizance of the detrimental effects of ETS on childhood illnesses that could potentially influence their health trajectory throughout adult life. A collaborative effort between communities, healthcare professionals and government bodies at all levels must be pursued to explore novel solutions within the realm of children’s environmental health. Thereby, successfully translating and communicating research findings into actionable interventions. Finally, the process of triangulation of evidence by means of reviews and pointing sources of bias in different study designs can help strengthen the degree of causality from multiple study designs [19].

The evaluation also focused on both the authors’ chosen keywords and those indexed in the papers. The commonly used term “human” was frequently observed, along with gender-specific words such as “male” and “female.“ Thus, when searching for papers related to ETS and CH, employing generic keywords may result in a more compendious search.

The conventional citation-based indicators do not assess the social media realm. As highlighted in additional file 3 the highest altmetirc score was displayed by a mechanistic review of tobacco smoke altering the immune responses in the lung triggering inflammation by Strzelak et al. (2018). This article was broadcasted through various news outlets and tweets; nineteen and seven times, respectively. On the contrary, the second article ‘Housing Characteristics and Children’s Respiratory Health in the Russian Federation’ published in 2004, was broadcasted by seven agencies but received low Twitter dissemination. From this study, we see the growth in Twitter and news outlets’ distribution of research cognates by a regress in blogs, CiteULike, and Facebook’s use to exchange scientific literature. Conjectures can be derived if these configurations demonstrate an alteration in the overall repute or if more distinct role changes amid social network types have led to this makeshift; however, further investigation is warranted. The percentage of papers with the maximal AAS suggests a huge diversity among the journals with 8% published in the Pediatrics journal followed by 5% each in the International Journal of Epidemiology and Environmental Research journal.

The relationship between the citations in WoS, Scopus, Google Scholar, and the observed AAS was poor. The lack of relation between the number of citations and AAS can be elucidated either by the varied nature of the items which have been taken for estimation or the distinct responses of a scholar/populace to a publication. A strong correlation was noted between Dimensions count and Scopus, WOS, and Google Scholar citation count. Dimensions count may be paramount since it can partially overcome the bias of Altmetric owing to the inconsistent features of social networks [20]. The AAS of environmental tobacco smoke and child health articles was not significantly correlated with the quartile of the journals. Similar results have been stated by other studies [21, 22]. Altmetric outcomes need to be conferred with prudence since the articles published before the burgeoning of the social media landscape may be under-represented [23]. Altmetrics evaluates the immediate influence of an article, in contrast to the traditional metrics where papers may take more than a decade to attain maximal citations [22]. Our findings displayed social media mentions reached a peak in the first five years after publication, this is in accordance with similar studies [21, 24].

Besides the aforementioned time delay in citations, the results of the study should be expounded with caution. Bibliometric and altmetric analyses have numerous inherent limitations. Firstly, landmark studies, over time, achieve fewer citations as their findings are absorbed into current knowledge without the necessity for referencing. This is referred to as “obliteration by incorporation” [16]. To mitigate this, we discerned articles by citation density. Second, self-citations can have an impact on citation counts. In this analysis, however, a major variance between the total number of citation counts and citations was not reported after excluding self-citations. Third, only articles published since July 2011 are picked up by Twitter. Also, the Bookmarklet works only on PubMed, arXiv, or Google Scholar pages containing a DOI [14]. Hence, the probability of influential articles not being cited by social media scientometrics cannot be ruled out. Fourth, altmetrics recognize the level of online activity of research without distinguishing between the publicity or the research output quality [15]. Fifth, altmetrics weight allocation in the generation of scores is related to the developer’s inference about their anticipated goal for every media platform [24]. Thus, there may be an imbalance in the contribution of diverse sources to AAS. Sixth, researchers can “game” the system by generating added mentions for their projects on a social forum [25]. This type of manipulation bias was improbable in the present study as Altmetric Explorer was used as a search engine.

Alternative metrics are in their early stages, and there is meager data about the elements of social platforms to certainly elucidate a definite association amidst novel metrics and bibliometrics. It is ambiguous if media presence leads to higher citations or if aspects that steer greater citation counts lead to increased social networking activity. Although the social web may have some cogency on the distribution of an article, alternative metrics ought to be employed alongside traditional bibliometric measures for assessing research impact comprehensively. Future investigations should explore methods to construct a comprehensive stratagem that integrates both citation-based and social media-based indicators for evaluating research outcomes.