Binge Drinking: The Top 100 Cited Papers

We conducted a review to analyze the 100 most-cited studies on binge drinking (BD) in the Web of Science (WoS) database to determine their current status and the aspects that require further attention. We carried out a retrospective bibliometric analysis in January 2021. The year of publication, authors, design, subject, journal, institution and lead author’s country, as well as the definition of BD, were extracted from the articles. The data on the country, year, thematic category of the journals and their rank were obtained from the Institute for Scientific Information (ISI) Journal Citation Reports 2020. The number of citations was collected from the WoS, and the h index was collected from the Scopus database. The citation density and Bradford’s law were calculated. The majority of the articles were empirical quantitative studies with a cross-sectional design published between 1992 and 2013 in 49 journals. There were 306 authors, mostly English-speaking and from the USA. The definitions used to describe BD are not homogeneous. The most-cited topics were the analysis of consequences, determinants and epidemiology. There is a need to unify the definitions of BD and base them on scientific evidence. The multidisciplinary nature of BD is not well reflected in each of the thematic areas discussed in this work.


Introduction
Binge drinking (BD) is a risky alcohol consumption pattern with a high prevalence at the international level, especially among the youngest population [1][2][3][4].
Given that BD is a pattern of consumption with multiple consequences at the organizational, psychological and social levels for both the binge drinker and those around him or her [5][6][7], in recent years, much research has been generated on this topic.
One method that allows the quantitative analysis of scientific production and the thematic evolution of a determined field of research is bibliometry [8]. Metrics such as the number of publications, the number of citations or the impact factor (IF) of journals are often used as measures of relevance or productivity [9,10].
Using such metrics, the analysis of the most-cited articles on a topic allows us to identify which contributions are more visible, which have been more recognized or those that have exerted greater influence on the scientific community beyond the limits of their field of expertise [11][12][13]. Furthermore, from the point of view of science policy based on excellence, examining highly cited works has been considered an option for the detection and monitoring of "excellent" scientific research [11].
In recent years, numerous studies have been conducted to determine and analyze the most-cited papers in various fields of knowledge. In the health fields, such analyses have been performed in areas such as surgery [14], cardiovascular disease [15] and ophthalmology [16]. Analyses have also been conducted in areas of psychological health, such as occupational stress [17] and obsessive-compulsive disorder [18]. However, this methodological approach has been limited in the area of addictions [19]. The relevant studies have focused on substance use disorders [20,21], and bibliometric analyses on BD are nonexistent.
The main objective of this study is the analysis of the 100 most-cited articles in the Institute for Scientific Information (ISI) Web of Science (WoS) database (Thomson Reuters, Philadelphia, PA, USA) to provide a unique perspective on the current situation and opportunities for BD research, intervention and formation in the coming years.

Search Methods
On 7 January 2021, the WoS Core Collection was searched with the keywords: "binge drinking", "heavy episodic drinking", "heavy drinking", "heavy sessional drinking", "dangerous drinking", "risky single-occasion drinking", "high-risk drinking", "risky single occasion drinking", "high risk drinking", "excessive episodic consumption", "frequent binge drinking", "concentrated drinking episode" or "episodic heavy drinking". These are all terms suggested by Cortés and Motos [1] in their review on how to define and measure BD. The search terms were included in the "topic" field (title, abstract, author's keywords and KeyWords Plus). No filters were used in the fields of language, time, human studies, territory, affiliations or availability.

Search Outcome
A total of 15,894 studies were obtained. A modified approach to the method used by Lim et al. [22] and other authors [13,23,24] who have investigated the most-cited articles in different areas of healthcare was used. Three experts with more than 10 years of experience in BD research with youth (M.-T.C.-T., J.-A.G.-C. and C.G.-Í.) independently selected and analyzed the 100 articles with the highest number of citations. The inclusion criterion was that BD be the main topic or one of the main research variables. The exclusion criteria were studies that (1) were not related to alcohol consumption, (2) focused on general alcohol consumption without specific reference to BD or (3) considered BD as a secondary variable in the study. Disagreements among experts, although not frequent, were resolved in a consensus meeting. The flow of information through the different phases of the review (PRISMA flow diagram) is shown in Figure 1. For articles with the same number of citations, the one with the highest citation density (total number of citations/years since publication) was classified first.

Quality Appraisal
An evaluation of the quality of the bibliographic sample was not performed since this study was a bibliometric analysis. As an objective indicator, we show the volume of citations in the WoS, which, although not an absolute measure of the quality of an article [25], is an index of the impact the article has exerted in the scientific community, with the assumption that high-quality research will result in a greater number of citations than lower-quality research [26].

Data Abstraction and Synthesis
Information was directly abstracted from the 100 articles for the year of publication, authors, design, subject, journal in which the article was published and the institution and country of the lead author. The design was classified into six categories [27]: (1) review studies, including systematic reviews and meta-analyses; (2) development and validation studies of instruments or scales; (3) cross-sectional studies, including follow-up studies using questionnaires, surveys or interviews; (4) papers using a qualitative methodology; (5) discussions on topics or methods; and (6) other types of studies that are not included in any of the other categories, e.g., longitudinal studies. The subject was also classified into six broad categories: (1) BD consequences, (2) determinants, (3) epidemiology, (4) intervention proposals, (5) consumption trajectories and (6) other, less frequent topics (evaluation, terminology discussions, typologies of consumers and psychiatric comorbidity). Finally, information was obtained on how BD was defined in each study according to three aspects [1]: amount of consumption, time interval and frequency.
From the ISI Journal Citation Reports 2020, data regarding the country of publication, the IF of the journals, its thematic category and its ranking in that category were obtained. The data on the number of citations for each study were obtained from the WoS, and the h index of the most-cited authors was obtained from the Scopus database.
Finally, the citation density of the articles and Bradford's law were calculated. The citation density, or the average of annual citations, allowed us to obtain an index of the relative impact of an article regardless of the year of publication [23,28]. Bradford's law [29,30] allowed the establishment of an objective measure of the weight of each journal in the top 100 by distributing the total number of journals into three productivity zones with a similar number of articles but a decreasing number of journals.

Most-Cited Articles, Citation Density and Temporal Distribution
The top 100 most-cited articles related to BD are presented in Table 1 in descending order according to the number of citations. Together, they total 34,908 citations (ranging from 1280 to 180). Only 30 articles qualified as "citation classics", i.e., having 400 or more citations [31].
All of the articles were published across a 21-year range, between 1992 and 2013 (shown in Figure 2

Publication Type, Main Topics, BD Definition and Target Population
Most of the BD papers were empirical studies (n = 73). Of these, 39 were crosssectional studies, including questionnaire and follow-up surveys or interviews, and 34 had a longitudinal design. Literature reviews were less frequent (n = 19). Of these, three were meta-analyses, nine were systematic reviews and seven reviews did not provide details on the methodology used to select the articles. Finally, of the remaining eight studies, two focused on the development or validation of an instrument or scale, one was a qualitative study and five were a discussion of a method or topic.
Regarding the topics of the 100 most important studies related to BD, the analysis of its consequences stands out, with 28 publications and 32.22% of the citations (n = 11,246). This topic also showed the greatest increase in citations since 1999 and began its decline very recently in 2016 (shown in Figure 3). The second most-cited topic (n = 27; 8784 citations, 25.17%) was the study of the variables or determinants underlying BD, with a stable and low volume of citations until 2004 and strong growth over the following 10 years. With respect to the works on epidemiology (n = 14) and intervention proposals (n = 12), they had practically the same number of citations (n = 4324, 12.39%; n = 4034, 11.56%, respectively). The works on intervention were initially the second most cited between 2003 and 2006, with a large decrease since 2015. Finally, the works on the analysis of changes in consumption trajectories according to the growth and maturity of the subjects evaluated (n = 10; 3405 citations, 9.75%) increased their citations between 1996 and 2008 and have since remained stable. The rest of the articles focused on topics such as BD assessment (n = 3; 1384 citations), concept analysis (n = 3; 1121 citations), determination of consumer typologies (n = 2; 428 citations) or aspects related to psychiatric comorbidity (n = 1; 281 citations).
An analysis of the definitions of BD used (shown in Figure 4) showed that the most frequent (n = 48) differentiated the amount of alcohol consumption by gender (five or more standard drinks for men and four or more for women). In contrast, approximately 25% of the studies used a generic definition (five or more standard drinks) that did not differentiate by gender. The other articles offered specific definitions, and commonalities among them cannot be established. Three articles did not include a definition of BD. Regarding the samples of these studies, interest in the university population stands out (n = 44). In addition, the USA population was the most frequently studied (n = 71).

Authors and Institutions
A total of 306 authors contributed to the 100 most-cited papers on BD. Five of the studies were single-authored [41,71,98,119,125], and 20 authors contributed to three or more articles (Table 2).  With regard to the authors' affiliation, only the author who appears first in the author list was considered. These 100 authors belonged to 59 institutions from seven different countries, with a greater participation from the USA (n = 62), and 3 institutions from countries where English is not the first language (Finland, Lebanon and Switzerland).
Sixteen institutions contributed more than one paper. The institution with the largest contribution was the Harvard School of Public Health (USA) (n = 16). The University of California (USA) and the University of Washington (USA) appear second with five papers each. The Centers for Disease Control and Prevention (USA), Syracuse University (USA), the University of North Carolina at Chapel Hill (USA), the University of Michigan (USA) and the Center for Addiction and Mental Health at the University of Toronto (Canada) contributed three papers each.

Journals in Which the Top 100 Articles Were Published
A total of 49 journals published the 100 most-cited articles on BD (Table 3) As for the Journal Citation Reports (JCR) categories in which these journals are indexed (they can be classified in more than one category), 42.55% are in categories related to medicine (general, psychiatry, pediatrics, cardiology, obstetrics, etc.). Sixteen (30.04%) are in categories related to psychology (clinical, developmental and multidisciplinary); fourteen (29.78%) are in the category of substance abuse; and 23.40% are in the category of public, environmental and occupational health. The rest of the journals (n = 6) are classified into the categories of biochemistry, behavioral sciences, demography, education, neurosciences, neuroimaging, pharmacology and sport sciences.
The journals' IFs ranged between 45.54 and 1.214 (X = 4.03), with the JAMA: Journal of the American Medical Association at the top. With respect to the JCR quartiles of these journals in their categories, 74.46% are located in Q1, 23.40% in Q2, 4% in Q3 and 1% in Q4. The rest (n = 2) have not had this value calculated since they were incorporated in their category this year. All but two of the journals are published in the USA (71.42%) or UK (24.48%).

Discussion
To our knowledge, this is the first bibliometric study to systematically identify and classify the top 100 most-cited papers in BD research. The results of our study provide various benefits for researchers and practitioners in BD. First, the results help new researchers understand the types of contributions, approaches, topics, populations and research methods applied in highly cited papers, enabling researchers to learn from them in writing higher-quality papers that will likely receive high numbers of citations. Second, such reviews help both veteran and new researchers identify topics with more visibility or impact so as to carry out more incremental research in those areas, as well as to identify topics on which more research should be conducted for greater visibility. Third, using such reviews, researchers and practitioners can identify the most frequently cited researchers and institutions with whom to collaborate, receive advice and training and so forth. Fourth, the results help practitioners identify research considered of higher scientific "excellence" in the specific field of BD and to use the techniques, programs or results reported in these studies.
The results show that these articles had 306 authors from 59 institutions in seven different countries. The country that contributed the most papers was the United States; moreover, most of the authors are from English-speaking institutions, as seen in similar studies [12,13,23,132]. Judging by these results, research in other countries that have this pattern of consumption needs greater visibility [19], as BD shows differences and similarities between countries that affect the generalization of the results [62].
In addition, the articles under study were published in 49 journals, of which only 14 belonged to the specific category on this topic of "substance abuse", with the bulk of the journals belonging to the categories of medicine, public health and psychology. This differentiation reaffirms the multidisciplinary nature of the most visible research on BD, which transcends its specialty. If we focus on these journals' IFs, as in other studies [13,133], the results confirm that the most-cited works on BD have been published mainly in journals located in the first and second quartiles of their categories. Furthermore, although the number of citations ranges between 1280 and 180, only 30 of these articles qualify as "citation classics" [31].
The topic most referred to in these works is that of specific medical consequences (e.g., cardiovascular effects, effects on the fetus and brain damage). The few works on the psychosocial consequences are often limited to listing them without exploring any in further detail.
The second most-cited topic is that of context-related BD determinants, specifically regarding the social norms of fellow drinkers or of the areas in which they live. Most are descriptive studies that do not estimate the weight or importance of these determinants in the development of BD and are also subject to uncontrolled cultural aspects. These characteristics make it difficult to generalize to other samples, other contexts or other countries.
The results of the review fail to reflect the multidisciplinary nature of BD in any topical area discussed in this work. In the studies on the consequences, physiological aspects predominate. The studies on determinants highlight those of a social nature. Regarding intervention, the most-cited articles are those of a psychosocial nature: brief motivational interventions, those focused on aspects of the consumer's personality and, to a lesser extent, those based on normative feedback or marketing.
Several of the authors indicate that three main parameters are referenced in the research to characterize the operational definitions of BD [1,58]: the amount of alcohol consumed, the duration of the consumption episode and the time interval at which the presence of BD is recorded. In the most-cited work on BD, it is confirmed, on the one hand, that few articles include the three parameters and, on the other hand, that there is no consensus in defining each of these. Similar results have been found in recent reviews of the concept of BD [134].
The most commonly used parameter is the amount of alcohol ingested. However, in one in four studies, the definition of BD implemented does not align with any that are agreed upon in the literature. Additionally, very few differentiate the amount of alcohol ingested according to gender, ignoring the metabolic aspects associated with this substance [135,136]. Regarding the duration-of-consumption parameter, there is a notable inaccuracy of the interval evaluated, the most widely used being "a single occasion". The low rigor with which these highly cited studies define the patterns of BD is striking. For BD, it is not only the large quantity of alcohol consumed that should be relevant but also that consumption is carried out in a short and quantifiable time interval, which could be indicated by the number of hours the person spent drinking [137].
The last of the parameters, the time interval of BD, accounts for the episodic and irregular nature of this consumption pattern. The time periods defined by many of these works are short and do not allow accounting for this discontinuity, such as "the past week" or "the last 30 days". Therefore, the diversity of the definitions found in the most-cited articles on BD, with special mention of nondifferentiation by gender or the differences in the time intervals of consumption, highlights the need to visualize conceptual advances in BD that will allow the homogenization of the results derived from the research.
As limitations of this work, first, we used only one database for our search for the most-cited papers. This procedure conditioned our list and ranking, which may have differed if we had used a different database, such as Scopus or Google Scholar [138]. Citations obtained exclusively from the WoS are limited to the sources selected using this database, which, for example, does not include citations in books and journals published in languages other than English. Nevertheless, we chose this database because, although it does not provide complete coverage, it has been extensive and multidisciplinary since 1900 and includes more than 12,000 journals with an impact across the world, including open access journals [139]. Second, self-citations and citations in lectures, textbooks and web-based literature have not been considered [140]. The positive or negative nature of the citations of the top 100 articles were not considered in this work, and thus, we were unable to determine the level of agreement, disagreement or criticism from the scientific community regarding the topics covered in these articles [141,142].

Conclusions
The analysis of the 100 most-cited articles in BD offers a unique perspective on the most visible works in this area, providing insights about the need to unify the definitions of binge drinking based on the scientific evidence agreed upon by experts. Moreover, the topical areas analyzed do not clearly reflect the multidisciplinary nature of the BD. In the same way, research from other countries that have this pattern of consumption needs greater visibility to show the differences and similarities between countries that will affect the generalization of the results.