A bibliometric analysis of drinking water research in Africa

.


INTRODUCTION
At the moment the priority research areas in Africa are concerned with easement of disease burden, mitigating water scarcity and enhancing of food security (Pouris and Ho, 2014).In particular, drinking water research continues to attract intense interest from African researchers (Dagdeviren and Robertson, 2011).This has lead to a surge in publication output in this area, especially in recent years.This is because it is generally believed that affordable access to quality drinking water could both mitigate food shortage and help control the disease burden among communities (Hunter et al., 2010).
African research output in drinking water is, therefore, diverse, ranging from such aspects as hydrogeochemistry (Ghiglieri et al., 2012), availability assessments (Reimann et al., 2003), quality assessments (Campbell et al., 2003), and treatment (Mameri et al., 1998) to policy and management issues ( Van der Zaag, 2005).To facilitate progress in water research in Africa, there is a need to consolidate the divergent knowledge base so as to focus Africa's research efforts in these disciplines.
Attempts towards this desired data integration include several systematic reviews by African research institutions (Kusangaya et al., 2014;Weinzierl and Schilling, 2013).While, systematic literature reviews are used in interrogating research status and often unravel useful details that could direct research and policy issues in the subsequent periods, such reviews address only a limited scope of data and are unhelpful in consolidating divergent data in open research fields like drinking water, and even more so if the field of study is multidisciplinary with authorship that traverses socio-political and academic borders.
It is found that in such cases bibliometric reviews provide more elaborate analyses (Li et al., 2011).The latter studies are based on quantitative statistical examination of publication trends, anchored on word-clusters and the distribution of terms in article titles, author keywords, and KeyWords Plus, among other indicators (Fu et al., 2013b;Khan and Ho, 2011).The criteria for bibliometric assessment is, thus, founded on examination of trends in research parameters, which adduce data used to assemble information that would characterize the evolutionary structure of research in a particular field of study (Feller and Bernoux, 2008).The analyses are geared towards exposing knowledge gaps and revealing emphases that would drive successive research in the study field (Nachega et al., 2012).
In view of the divergent geophysical characteristics of Africa, bibliometric studies that would address regional disparities in primary research areas are desired for focusing research efforts in Africa toward specific socio-economic challenges that face African communities (Chuang et al., 2011).This would provide an all-inclusive view of available data and inform tailored interventions.To date, only a few bibliometric studies have been conducted to appraise research work in the African context (Chuang et al., 2011;Pouris and Ho, 2014).
The current work was designed to screen drinking water research in Africa during the years spanning 1991 to 2013.The bibliometric research data were based on Science Citation Index Expanded (SCI-Expanded) in the Web of Science Core Collection.Trends in research productivity, author contributions and international collaboration were analysed.It was anticipated that the results would provide insight into the existing research trends and help researchers to discern the direction of science in drinking water research in Africa.It is hoped this will in turn help researchers to identify and engage knowledge gaps towards informed interventions into the problems that face the African continent.

METHODS
The data were drawn from the online version of SCI-Expanded, in Thomson Reuters Web of Science (updated on 30 July 2014).Algeria, Angola, Benin, Botswana, 'Burkina Faso' , Burundi,

Main document type
A total of 1 719 documents were found within 9 document types indexed in the Thomson Reuters Web of Science.The most common document type was the research article (95% of the 1 719 documents), followed by proceedings papers (8.3%), review papers (3.2%), meeting abstracts (0.70%), notes (0.47%), letters (0.47%), editorial materials (0.12%), a correction, and a book chapter.The patterns in distribution of the document types were comparable to those reported in literature (Chuang et al., 2011).However, there were no bibliographical items, bibliographies, database reviews, research discussions, software reviews and reprints (as found in studies by Fu et al., 2013a;Tan et al., 2013) found in this study.
Citation data for different document types were analysed (Table 1).The total citation, TC 2013 , for each document type was consistent with the total publication output (TP) within the document types.This shows that higher publication output of a particular document type led to higher visibility and likelihood of citation within that document type.It was not surprising that reviews had higher citations per paper (CPP) than articles because review papers cover a wider perspective of a research topic and present information to readers in a more integrated manner, which help readers to get a broader perspective of the subject matter compared to other primary document types.For this reason, reviews receive more citations than articles (Guimaraes and Carlini, 2004).It was, however, unusual that conference proceedings in the current analyses had higher CPP than the articles.This indicates that drinking water research in Africa is still developing as researchers in mature research fields generally consult articles from reputable sources known for their high standards and rigorous peer-review.However, only the 1 633 original research articles found in the current work were further examined as articles represented the majority of the peerreviewed documents in this field.

Publication output and language
Trends in distribution of research articles by year of publication were analysed (Fig. 1).The annual output of research articles increased from about 20 articles in the early 1990s to 71 articles in the year 2004.Publication output decreased to 54 articles in 2007 before again increasing to 194 articles by the year 2013.The observed expansion in the research output reflected increased use of information technologies (Samuels et al., 1991) such as computers, the Internet (Blau et al., 2003) and mobile phones (Ailamaki et al., 2003), which enhanced data handling and transmission and access to scientific information during this period.The period under study also coincided with the time of rapid expansion in higher education in Africa leading to increased research output from the continent (Varghee, 2004).
The graph of 'number of articles' versus 'time in years' depicted in Fig. 1 is convex, which affirms that drinking water research in Africa is a developing study field.Similar plots normally assume a characteristic sigmoid shape for maturing study fields (Yu et al., 2012).
All in all, English was the language of choice for communicating scientific data, having been used to publish 1 582 or 97% of the 1 633 research articles., followed by French (2.9%), German (0.12%) and Spanish (0.061%).Scientific language use reflected historical ties between Africa and Western Europe, which are traceable to the 19 th century 'European scramble for Africa' (Griffiths, 1986).The publication languages found in the current study were, therefore, fewer than is normally the case in similar analyses (Fu et al., 2013a;Mao et al., 2010).

Article characteristics
As depicted in Table 2, drinking water research in Africa became increasingly collaborative over the period studied.The mean number of authors per article increased from 2.7 to 4.8 and mean paper length rose from 6.3 to 10 pages between the early 1990s and 2013.The number of bibliographic sources per articles also increased from 13-22 citations, which was typical of research articles in the early 1990s, to 40−45 citations per article after the year 2010.

Research focuses of drinking water research in Africa
Statistical analysis of the frequency of key terms in article titles, authors' keywords, and KeyWords Plus provides details of research progress and indicates the direction of science in a research field.Distribution of title words, authors' keywords, and KeyWords Plus in different periods can be used to determine research focus (Li et al., 2009) and word clusters have been applied to characterize articles in the subsequent years (Mao et al., 2010).
Accordingly, the most frequent KeyWords Plus terms were more varied than the top title words but less diverse than the top authors' keywords.The term 'water' appeared in the top 10 category across title words, authors' keywords and KeyWords Plus.Two terms, 'drinking water' and 'rats' , were found among the top 10 for title words and authors' keywords.Concurrent appearance of the terms 'South' (141; 8.6%) and ' Africa' (124; 7.6%) in title category alludes to the name 'South Africa' , indicating the national origin of the research work.This underscored the central role played by South African research institutions in water research in Africa during this period.The term 'oxidative stress' , which relates to the effect of reactive nitrogen oxides to biological systems (Betteridge, 2000), appeared in the top 10 most used

Figure 1 Number of articles and citations per article by year
terms across authors' keywords and KeyWords Plus.This suggests a significant interface between environmental toxicology, ecology and drinking water research in Africa during this period.
Besides the overarching regional imbalance in water research, availability of research capital including research technologies and related structures controlled inter-state disparities in drinking water research involvement by African researchers.Consequently, South Africa, Egypt and Nigeria (Chuang et al., 2011;Nachega et al., 2012;Pouris and Ho, 2014), which are among the countries with some of the most established research structures in Africa, dominate research output across all study fields.

Research collaborations
Collaborations among researchers, research institutions and countries play a pivotal role in contemporary science.To evaluate exchange of scientific information between Africa and the rest of the world, research collaborations between African scientists and research institutions from outside of Africa were analysed.The top 10 most collaborative institutions were ranked by total number of publications (TP), including correspondingauthor articles and first-author articles.A total of 799 or 48.9% of all drinking water research articles were published by African research institutions in collaboration with institutions from outside Africa.Over 26% of these articles were attributed to the 27 most collaborative institutions from outside Africa with at least 5 co-authored research articles with African researchers in this period.The top 5 institutes in this category were Centre for Disease Control & Prevention (30; 1.8%); King Saud University (11; 0.67%); University of Montpellier (11; 0.67%); King Abdulaziz University (11; 0.67%); and University of Bergen (10; 0.61%).The list of the top 27 collaborative institutions from outside Africa was dominated by institutions from the UK (7; 26%), France (5; 19%), USA (4; 15%), Saudi Arabia (4; 15%); Germany (2; 7.4%), and Norway (2; 7.4%).Ireland, Japan and Spain were each represented by one institution among these 27 elite institutions.
The collaborative agencies that retained first-authorship of collaborative articles also retained corresponding-authorship (r 2 > 0.988).German (92%), Saudi (72%), and Norwegian (71%) institutions retained highest proportion of first-and corresponding-authorship in their collaborative articles with African institutes, whereas French (32.7%) and Spanish (38.7%) agencies had the lowest proportion of first-author and corresponding-author articles with African researchers.

Article visibility and citation trends
To assess the visibility of research articles, the number of times an article was cited from publication till the end of 2013 (TC 2013 ) was used as an indicator (Chuang et al., 2007).Reader preferences and scientific impact of published work were studied by analysing the 19 most cited publications in drinking water research in Africa from 1991 to 2013.The list of the most cited articles (TC 2013 >60) is given in Table 5.With the single exception of one review paper, the entire list of top cited works comprised of research articles.The main subject areas covered by the 19 most cited articles were: 'water defluoridation' (4 articles) and 'water fluoride' (1 article), which were cited 572 times; followed by 'in vivo animal tests' (4 articles), cited 372 times; 'solar water purification technology' (3 articles) with 343 citations; 'general water purification technologies' (2 articles) with 142 citations; and 'heavy metal water pollution' (1 article) with 89 citations.
The high readership of articles on the subject of water fluoride was expected when it is considered that large portions of Africa, including regions along the Great Rift Valley from Sudan through Southern Sudan, Djibouti, Eritrea, Ethiopia, Kenya, Uganda, Tanzania, Malawi, down to South Africa; West African states of Niger, Nigeria, and Senegal (Ayoob and Gupta, 2006); and Northern Africa states of Egypt (El-Sadaawy and El-Said, 2014), Algeria (Achour andYoucef, 2009), andTunisia (Hamdi andSrasra, 2009) are well known high-fluoride areas of the world.Therefore, in view of the high toxicities and the risk of high-fluoride exposure for the communities in the affected areas, the fate of fluoride in groundwater resources remains a major public health concern (Wambu et al., 2014).
Figure 2 shows that the visibility of the articles increased rapidly after the year 2000.Whereas the apparent reeason for this upsurge was not immediately clear, the current study period coincided with a time of rejuvenation and expansion in higher education and the academic sector in Africa (Varghee, 2004).The period also corresponds to the inception of the computer era and expanded use by researchers in Africa of electronic data retrieval systems and the Internet to access research databases (Ailamaki et al., 2003;Blau et al., 2003).An analysis of citation trends of drinking water research articles among the top African countries is depicted in Fig. 3. Visibility of articles reached its maximum within 2-3 years of publication and decreased thereafter to near-zero in 17−22 years, as previously reported by Chuang et al. (2007) and Zhang (2012).However, the actual duration of optimum scientific visibility of articles varied from country to country.Articles from Tunisia had the highest mean visibility of 2.0 citations per article.They were followed by articles of Egyptian (~1.5 citations), South African (~1.4 citations), and Nigerian (~1.0 citations) origin.As can be seen from Fig. 4, internationally collaborative articles had the highest visibility and scientific impact followed by inter-institutional collaborative articles, single-country articles and single-author articles, respectively.

CONCLUSIONS
An overview of drinking water research in Africa during the 1991-2013 period has been presented with respect to annual publication outputs, Web of Science subject categories, research areas, journals, institutions, countries, and trends in collaboration, research emphases and tendencies.The publication output in this field increased at exponential rates over the entire study period.It has been adduced therefore that drinking water research in Africa is a developing field of study.The key subject categories of study were: 'water resources' , 'environmental science' , and 'environmental and occupational health public' .Environmental toxicology and ecology, therefore, significantly controlled Africa's water research during this time.Research output by African countries was, however, controlled by availability of research capital and regional differences in environmental factors.The overarching research challenges were disease burden in the sub-Saharan region and water shortage in the arid north.About 25% of all publications in drinking water research were published by South African institutions and the distribution of drinking water research by region for the rest of Africa followed the order: North Africa > West Africa > East Africa > Southern Africa > Horn of Africa > Central Africa.Collaborative research became increasingly attractive and about half of all articles were co-authored by African researchers and collaborative institutions from Europe, North America, and the Far East.The most collaborative countries were France and USA.The highest visibility and scientific impact of publications were associated with review articles and internationally collaborative articles.

RECOMMENDATIONS
Drinking water research in Africa bore the marks of a developing research field and there is still a paucity of data in many aspects of this field.The African states and international community must make deliberate efforts to promote research in this field and bridge the gaps to meet the technological demand for research data in this area.Yousef et al. (2003) 12 ( 76) Thymoquinone attenuates ifosfamide-induced Fanconi syndrome in rats and enhances its antitumor activity in mice Badary (1999) 13 ( 72) Antioxidant effect of vitamin E and selenium on lipid peroxidation, enzyme activities and biochemical parameters in rats exposed to aluminium El-Demerdash ( 2004 An interesting topic for future research would be to break down the analysis of the data by African regional blocks in order to examine the type of water research that dominates in the different regions of the continent.
Current results showed that environmental safety is increasingly becoming a subject of concern to Africa.Deliberate efforts must be made by all stakeholders and players in the relevant sectors to safeguard African water resources, especially the crossborder water resources, to curb any further degradation.

Figure 2 Figure 3
Figure 2 Number of citations per article by year for the top seven most cited articles in drinking water research in Africa from 1991 to 2013

Figure 4
Figure 4 Effect of collaborations on article visibility

TABLE 1 Distribution of water research by document types of publi- cation output in Africa between 1991 and 2013
TP: number of articles; TC 2013 : number of citations since publication to the end of 2013; CPP: citations per publication.

TABLE 2 Characteristics of drinking water scientific articles from Africa between 1991 and 2013
TP: number of articles; AU: number of authors; NR: number of cited references; PG: number of pages; CP: number of internationally collaborative articles; %: percentage of internationally collaborative articles in each year

TABLE 4 Top 21 African countries with more than 10 articles
TP: total articles; TPR (%): rank and the percentage of total articles; IPR (%): rank and the percentage of independent articles; CPR (%): rank and the percentage of international collaborative articles; FPR (%): rank and the percentage of first author articles; RPP (%): rank and the percentage of the corresponding-authored articles; SPR (%): rank and the percentage of single-author articles; N/A: not available.http://dx.doi.org/10.4314/wsa.v42i4.12Available on website http://www.wrc.org.zaISSN 1816-7950 (Online) = Water SA Vol.42 No. 4 October 2016 Published under a Creative Commons Attribution Licence 617

TABLE 5 Top 19 articles with TC2013> 60
Effect of ascorbic acid and Vitamin E supplementation on semen quality and biochemical parameters of male rabbits