Background The field of bioethics is constantly evolving. To investigate trends in the field of bioethics, we conducted a quantitative analysis of the top-cited articles in bioethical journals over the past 40 years.
Methods Retrospective quantitative study of the 20 most cited bioethics articles published each year from 1975 to 2014 were conducted. Article samples were selected from a list of the most relevant 100 journals in the field of bioethics.
Results In total, 800 top-cited articles between 1975 and 2014 in the domain of bioethics were retrieved and analysed. More than half of them were composed by single authors, but multiauthorship became more prevalent with time. The majority (84.5%) of these highly cited articles originated from the USA (65.3%), UK or Canada, though the proportion of other countries increased in recent years. Almost half (44.6%) of the highly cited articles belonged to the subfield of clinical ethics, but other subfields such as research ethics, public health ethics and neuroethics became more prominent. Overall, the distribution of Thesaurus keywords and subfields became more diverse over time, and the number of journals publishing top-cited articles doubled. Furthermore, the empirical ethics approach increased over time in our sample of top-cited articles.
Conclusions In sum, the forefront of bioethics is getting more diversified, collaborative and international. The presumed ‘mainstream’ becomes less dominant over time, as more highly cited articles come from new subfields, discuss new topics, use more Bioethics Thesaurus keywords, more authors participate and more countries other than the USA contribute to bioethics journals.
- History of Health Ethics/Bioethics
Statistics from Altmetric.com
Being multidisciplinary on one hand and pluralistic on the other, bioethics is not easily defined and is a diffuse body of scholarship.1 It assembles contributions from various disciplines2 and covers diversified research areas.3 The field of bioethics has been labelled as ‘multidisciplinary’, ‘interprofessional’ and ‘multicultural’, while some interpret these labels as a cover-up of its disunity and fragmentation.1 ,3 We see bioethics as ethical inquiries into phenomena or problems of biomedicine, public health, environment and other areas relevant to human health and life. It is an inclusive research domain or a field of inquiry still in the process of defining itself. Nonetheless, our rudimentary understanding cannot contribute to the ongoing debate on the nature and scope of bioethics. For that purpose, we conducted a bibliometric analysis on the most influential work in bioethics journals to see how this dynamic field evolved over time.
Previous bibliometric research focusing on the field of bioethics has addressed the emergence of the empirical approach,4 time variation of specific topics5 ,6 and authorship.7–9 These studies provide a general picture of bioethics from different perspectives; yet none has yielded a focus analysis of bioethics literature, as has been done by the classic citation studies in many other disciplines.10–12 Citation count is a quantitative proxy for the significance and influence of journals, scholars and article topics.13 In this study, we collect and analyse the top-cited articles published in bioethics journals to identify the major evolvements in this field. We are well aware that these highly cited articles do not comprise a comprehensive profile of bioethics but rather are only a sample of the entire corpus of published works. Nonetheless, we believe that documenting such influential work in bioethics can provide unprecedented information for the progress of this flourishing research domain and the identification of its major intellectual advances and knowledge producers, and inspire researchers to produce equivalent or superior scholarship.
Our research concerns how the influential work of bioethics, represented by the top-cited articles published in bioethics journals, has developed during the past 40 years. Specifically, we are interested in the following questions:
What is the trend of authorship and geographical allocation of these highly cited articles?
What themes received the most attention, and do popular themes change with time?
Do these highly cited articles manifest the ‘empirical turn’14 in bioethics?
What are the characteristics of the journals that publish highly cited articles?
Subject matters in the field of bioethics have not been delineated in a clear and consistent way. In this study, we take a conventional approach using bioethics journal as the primary criterion for defining ‘bioethics’, which has been applied in previous bibliometric research.4 ,9 The strategy of using a journal as an identifier attempts to avoid reliance of full-text indexing that would arise from using a textbook definition of bioethics as a search strategy or only relying on a subjective analysis of examining the methodology or the identity of the author.
We compiled a list of bioethics journals to analyse, using 100 journals selected from a previously published list based on the impact factor (see also online supplement III).i We created a dataset of articles using the tool Publish or Perish (PoP) to search by journal title to retrieve 51 587 article metadata records and their citation counts from Google Scholar from 2 to 17 April 2015. To obtain a balanced sampling of bioethics articles over time, we adopted a snapshot approach that evenly identifies the top-cited articles through a time frame of 1 year. Therefore, for each of the 40 years from 1975 to 2014, we selected the 20 cumulatively most cited articles published in that year to identify an initial sample of 800 articles (as retrieved from Google Scholar on 17 April 2015). However, these 100 journals do not always publish bioethics articles in the 40-year time span, and PoP sometimes retrieved non-journal articles; we therefore established a set of inclusion and exclusion criteria to help improve our sample to more accurately reflect our understanding of bioethics. Eligible articles have to fulfil the following criteria: (1) peer-reviewed journal articles; (2) the title and abstract indicate an ethical inquiry of questions on health-related contexts (see more details in online supplement III). Both authors used the criteria to examine the sample separately, discussed and reached a consensus on the final sample set (see online supplement I).
We have carefully selected a series of indicators to describe the sample for further analysis. Indicators based on basic information include: year of publication, number of authors, first author's institution and location, and journal. Indicators relevant to content include: five top-frequency Thesaurus keywords appearing in full text, subfield and orientation. Thesaurus keywords (keywords afterwards) were extracted from full text using the online Automatic Keyword Indexing Tool based on the Bioethics Research Library's Bioethics Thesaurusii. Subfield indicators were developed based on previous taxonomic literature15–17 and personal experience and understanding of what constitutes the field of bioethics. The 10 categories of subfield are: general bioethics, clinical ethics, research ethics, biotech ethics, public health ethics, neuroethics, professional ethics, law and bioethics, religious bioethics and environmental ethics (see online supplement II table S1). Orientation indicators have two categories: empirical approach refers to a descriptive ethics research approach that applies social science methods; conventional approach refers to normative ethics or meta-ethics research that does not rely on empirical methods. We also tracked the variation of some topics such as AIDS and enhancement. To facilitate the identification of the temporal pattern, we grouped data in 10 years and the analysis is done by decade. Except for keywords that were extracted from full text automatically, other content-relevant indicators were assigned manually by one author based on title, abstract and full text if necessary. The assignment of indicators was double-checked by the second author afterwards to minimise personal bias. Data were analysed in Microsoft Excel 2013 using descriptive methods.
In total, 800 top-cited articles between 1975 and 2014 in the domain of bioethics were retrieved and analysed. According to Google Scholar (17 April 2015), these articles were cited 73 325 times altogether, with an average 91.66 (SD 76.31, median 79.5, Q1=42, Q3=117iii) accumulated cites per article (see online supplement II figure S1). The two articles having the highest cumulative citation count and highest citation count per year post publication are ‘Quality of life in cancer patients: an hypothesis’18 (1007 in total, 32.48 cites/year) published in the Journal of Medical Ethics in 1984 and ‘Managing incidental findings in human subjects research: analysis and recommendations’19 (366 in total, 52.29 cites/year) published in the Journal of Law, Medicine and Ethics in 2008. This latter article also had the most contributors in our sample, with 21 authors listed.
Authorship: number of authors, institutions and their geographical locations
A trend of increasing multiauthorship of highly cited articles has emerged and continued rapidly (figure 1). On average, 1.88 (SD 1.55, median 1, Q1=1, Q3=2) authors contribute to one article, though more than half (57.9%) were written by a single author. The mean number of authors has increased from 1.35 (SD 0.73, median 1, Q1=1, Q3=1) in 1975–1984 to 2.53 (SD 2.12, median 2, Q1=1, Q3=3) in 2005–2014. The prevalence of single-author articles in our sample decreased from 75.5% in the first decade to 38% recently.
We categorised the sample's geographical distribution according to the location of the first author's institution affiliation. The articles in our sample come from 31 countries: 71.3% of articles were published by authors affiliated with institutions from North America and 23.4% from Europe. Although the USA remains the most productive country, the geographical origin of these articles is becoming more international over time (figure 2). In the first decade, only eight countries published highly cited articles, which increased later to 14 countries in the second and 20 countries in the recent two decades.
We also tracked the first author's institution and institution type (see online supplement II table S3, figures S2 and S3). In total, 370 institutions were identified in our sample, of which 82.9% were published by scholars working in 269 academic universities. Universities remain the major author-affiliated institutions, while non-profit organisations, companies, research institutes, hospitals and government agencies publish a smaller but significant portion of highly cited articles in bioethics.
Content: Thesaurus and subfield
To investigate the themes of highly cited bioethics literature, we adopted complementary objective and subjective approaches. The objective Thesaurus analysis automatically identifies the five most frequently appearing keywords from each article's full text. We identified 378 distinct keywords in the article text, and on average, each keyword was identified in 10.58 (SD 30.23) articles. The number of unique keywords identified in later decades is generally getting more diverse than earlier decades in our sample: from 179 (1975–1984) to 195 (1985–1994), 222 (1995–2004) and 212 (2005–2014). The most frequent keywords identified were: health (371 articles), patients (222), ethics (183), research (172), life (153), medicine (124), healthcare (121), physicians (108), law (94) and consent (82). Some keywords frequently appear in text with other keywords: consent, for example, appears most frequently alongside research (51 articles) or patients (27), while research appears most often with health (78) or ethics (58), and philosophy is found with medicine (11) or life (9).
We manually categorised the articles into 10 different subfields based on their title, abstracts and for some, full text. Almost half of them (44.6%) belonged to clinical ethics, while research ethics (12%), law and bioethics (9%) and public health ethics (9%) also occupied a significant place. Judging from the volume of highly cited articles in the subfields, the dominant place of clinical ethics is declining (figure 3). The subfields of research ethics and biotech ethics had a significant increase in the last two decades, while in the last 10 years, neuroethics, public health ethics and professional ethics emerged.
These articles cover a wide range of topics in bioethics. Some issues, such as end-of-life issues, patient decision making and ethics education are likely to appear in all four decades (see online supplement II tables S2 and S5). Some specific topics attracted significant attention in a certain time period and were less discussed in other decades, such as discussion on organ transplantation and AIDS in 1985–1994, and discussion on enhancement and moral distress in 2005–2014.iv
Method: emergence of empirical ethics
We investigated the orientation of our sample by examining whether social science methodologies were applied in collecting and analysing empirical evidence on normative issues. Despite the fact that the empirical approach accounted for only 16% of the sample, its proportion increased slowly with time, from 8.5% in the first decade to 23.5% recently (figure 4).
The average number of authors listed per article applying an empirical design is 3.15 (SD 1.93), and that of articles with non-empirical design is 1.64 (SD 1.34). Most empirical research was conducted in the subfields of clinical ethics (50%), research ethics (15.6%) and professional ethics (15.6%). The distributions of non-empirical research in subfields are the following: clinical ethics (43.6%), research ethics (11.3%), law and bioethics (9.8%) and other (35.3%).
Out of the 128 articles that adopted an empirical approach, the most common data sources were medical professionals (34.4%) and patients or family members (18.8%). Frequently appearing topics included end-of-life issues (11 articles), moral distress (8) and ethics education (7) among others.
Journals: characteristics of the source of top-cited articles
Out of the bioethics journals we checked, including 15 non-English journals, only 44 of them have published top-cited articles in the last 40 years, and all are English-language journals. Out of these 44 journals, 10 have published 83.4% of the top-cited articles. Furthermore, our findings suggest that journals may have their ‘preference’ on authorship pattern and orientation concerning the highly cited articles (see table 1).
The number of different journals publishing highly cited articles also shows a trend of increasing diversity. From 1975 to 1984, only 14 bioethics journals published highly cited articles; that number increased to 15 from 1985 to 1994, 26 from 1995 to 2004 and 28 from 2005 to 2014. While our analysis indicates that some journals publish highly cited articles regularly, it also shows that new journals are publishing more and more highly cited articles with time. The distribution of highly cited articles among journals seems to be decentralising as bioethics develops over time.
This study intends to show the highlights of the field of bioethics in the last four decades by analysing 800 top-cited articles published in journals closely relevant to bioethics. It is noteworthy that a high citation count does not necessarily indicate high-quality scholarship in bioethics.20 Nonetheless, we believe these highly cited articles are more likely to have greater ‘impact’ and higher research significance to bioethics. We may therefore have a higher chance to find the topics among these articles which can later develop into a major advance of the field.13 Although our findings mainly concern the characteristics of articles with high citation counts, a great proportion of them is consistent with previous research in the field of bioethics in general.
Decline of single authorship and diversification of geographical knowledge producers
Although the pattern of single-author publication dominated the early years of bioethics, it is less the case now. The phenomenon of multiauthorship became unprecedentedly prevalent in the field of medicine and bioethics.8 Such growth may either be a positive sign of increasing cooperation and fair credit distribution among scholars,7 or a negative ‘authorship inflation’21 that indicates honorary or other inappropriate authorship.7 We cannot be certain which case fits more of our findings that highly cited articles manifest the same tendency of a fast growth of multiauthorship. Notably, both the average number of authors (1.88) of highly cited articles from 1984 to 2014 and that from 2005 to 2014 (2.53) are higher than that of bioethics in general (1.52 from 1990 to 2003).8 This could indicate that either greater cooperation or more honorary authorship or both exist in highly cited articles than average bioethics articles. The reasons behind this are complex and multifold, including more nuanced and robust scholarship through collaboration, multidisciplinary interest in the publication, and broader social network promotion and citation.22
The USA, the UK and Canada are the most fertile countries in bioethics,9 and they surpass other countries in producing highly cited articles. Over time, the geographical variation is increasing, while the dominance of the USA is declining, which also corresponds to the trend in the whole field of bioethics.9 The lack of representation of low/middle-income countries in bioethics becomes magnified in highly cited articles, with <1% of them coming from middle/low-income countries. To overcome this geographical imbalance, more research and author contributions from low/middle-income countries in bioethics must be supported. Researchers from low/middle-income countries must seek increased international communication with peers in bioethics.
In short, our findings indicate that highly cited bioethics articles are more often single-authored, university-produced and published by authors affiliated with institutions located in the USA. However, the trend is moving towards multiauthorship and becoming more internationalised, even if dominated by developed countries. We see no reason why this trend will not continue.
Changing themes of bioethics
The forefront of bioethics varied over time;6 each period can be defined by topics or areas of interest that attract a great deal of discussion23 and thus citations, and some themes are more enduring than others. Some variation of themes among these highly cited articles can be explained mainly by what happened in the general field. Reasons applicable to the general field, such as reaching a consensus or saturated discussion leads to decline and emerging social issues raise the discussion volume,6 can explain the trend of highly cited articles. For example, the discussion of AIDS in highly cited articles synchronise well with that in general bioethics;6 both had a peak in 1985–1994 and a significant decline afterwards. However, other factors relevant to citation counts, including the prestige, internationality and visibility of the journal, scientific utility, quality and accessibility of the article, number, gender and nationality of the author,24 may also play a role.
Although there used to be or still are some dominant keywords or subfields, such strong dominance is declining, and the themes of highly cited articles are getting diversified as more keywords appear and minor subfields fill a larger proportion of highly cited articles in bioethics. We see this trend happening in the example of health as a frequent keyword and clinical ethics as a prevailing subfield; their dominant position is weakened by the increase of other keywords such as ethics, research, nurses and industry and the expansion of subfields such as research ethics and neuroethics.
The complementary approaches of keyword analysis and subfield analysis discovered similar variations of the themes, and most findings from both approaches correspond well. However, it would be a mistake to replace one by another. On one hand, one frequent keyword can tell little about the article since one article may have many keywords, and keyword identification does not include the context of their usage. Although some articles may be categorised by two or more subfields, it is possible to use one main subfield as a snapshot of the article. On the other hand, Bioethics Thesaurus keywords are free from personal bias, and can provide extra information about the full text.
Emerging empirical approach
Our findings confirm that the impact of the empirical approach is increasing among highly cited articles, though the conventional approach remains most common and influential. A similar trend is observed in the general field of bioethics,4 ,25 which explains the expansion of empirical research in highly cited articles.
Highly cited articles with an empirical orientation have more than twice the number of authors than those with a conventional orientation. They favour subfields such as research ethics, professional ethics and topics such as moral distress and ethics education that are prominent only in recent decades. This expansion of empirical approach may therefore partly explain the growth of highly cited articles in these subfields and topics. Our findings on highly cited empirical articles concerning authorship, topic and sample population are all consistent with that in the general field of medicine and bioethics.4
More knowledge producers in the field
Not all journals published highly cited bioethics articles, as in other fields.10 ,11 While our sample included articles published in 44 different journals, 10 influential journals published more than 80% of the articles in the sample. These 10 journals are not homogeneous in authorship and orientation of the highly cited articles that they published, which may be influenced by the profile, targeted audience and overall impact of the journal. Moreover, as few bioethics journals existed in the 1970s and 1980s, even less of them published highly cited articles; journals publishing highly cited articles double their number in the last decade, which may be explained by the growing number of bioethics journals and the rising influence of new or previously unrecognised journals.
Our sampling strategy may not be ideal due to the limitations of using journal as the primary criterion. First, journals not included in our sample also publish highly cited articles covering issues in bioethics. For example, New England Journal of Medicine published ‘Equipoise and the ethics of clinical research’ in 1987 that has been cited 1726 times, and JAMA published ‘What makes clinical research ethical?’ in 2000, with a citation count of 1656 (as retrieved from Google Scholar on 3 August 2016). However, identifying these articles will increase the complexity of creating additional search strategies and raise the challenge of examining the full text. Though our sample based on 100 journals may not cover the comprehensive range of this field, it can partly reflect the dialogue and discussion within the bioethics research community. Furthermore, bioethics journals in our list do not always publish bioethics articles, and our secondary screening may introduce some personal bias. Nonetheless, we used predefined criterion to examine the sample, documented excluded articles and reasons for exclusion, and triangulated results between two authors to make sure articles selected are relevant to bioethics, either core or peripheral. Future research may use other criteria to identify bioethics literature and compare results with ours.
Another minor limitation is that the citation count data used to select the 800 articles in our sample are generated by Google Scholar, while most other studies doing a similar research methodology used Web of Science to calculate citations. Since the calculation method is different between the two systems,26 our findings cannot be easily compared with these articles. Although Google Scholar is believed to provide broader citation coverage than Web of Science in the humanities,27 citation analysis articles based on Google Scholar remains difficult to locate which makes our findings less connected to the current literature. One way to overcome this problem is to perform similar research using Web of Science, which will be a possible direction for future research.
Our analysis of 800 top-cited articles epitomises the general field, since most findings are consistent with previous research. We have shown that the forefront of bioethics, perhaps as well as the general field, is becoming more collaborative, more internationalised, more diversified and decentralised.
This paper presents a condensed history documenting the trends of top-cited bioethics literature from 1975 to 2014, which enriches our understanding of its diversified and prominent scholarship. Future research may focus on explaining what determines why certain articles receive more citations, in the hope of motivating scholars in the field of bioethics to become more productive and influential.
The authors thank the two anonymous reviewers for their constructive comments and suggestions.
Contributors This study is conceptualised by both authors together. MH retrieved the original sample. MH and PJ performed sample screening separately, discussed and reached an agreement on the final sample. Both authors participated in data analysis, drafting the original manuscript and modifying the manuscript according to the comments of the editors and reviewers. Both authors reviewed and approved the revised draft.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement We agree to share our sample articles’ metadata to all interested readers. They will be found in our online supplementary files.
↵i Top 100 Bioethics Journals in the World. 2015. https://bioethics.georgetown.edu/2015/04/top-100-bioethics-journals-in-the-world/ Accessed: April, 2015.
↵ii Bioethics Thesaurus Automatic Keyword Indexing Tool. https://bioethics.georgetown.edu/library-materials/bioethics-research-library-databases/bioethics-thesaurus-database/
↵iii Because of the skewed data distribution for citation counts and authorship, we provide the first quartile (Q1), median and the third quartile (Q3), in addition to SD.
↵iv However, this does not mean such topics were not discussed in other time periods. Our findings only apply to the sample of 800 highly cited articles.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.