Annals of Library Science and Documentation 1988, 35(1), 34-42

BIBLIOMETRIC STUDY OF AIDS LITERATURE, 1981-84

A bibliography comprising 316 items on AIDSfor the period, 1981-84 was analysed to studythe characteristics of AIDS literature. The rateof growth has been found to be exponentialafter the first publication appeared in 1981. Alarge number of letters to the editors, compri-sing a total of 83 is a special observation made inthe study. 671 authors have contributed 261items, of which 78 items have been contributedby single authors. AIDS research has been foundhighly collaborative. In all 22 journals havepublished 272 items, (173 journal articles). Acore of 5 journals has been identified for AIDSliterature. A new journal - AIDS Research hasbeen founded in late 1983. Bradford's law didnot apply to the journal productivity distribu-tion while Lotka's law with a = 3.2 was foundto be applicable to author productivity distri-bution.

INTRODUCTION

Information as a resource for development, as acatalyst for scientific and innovative activitieshas been recognised now by the sociologists,economists, scientists, technologists and allgroups of potential users of information. Insuch a situation, it becomes necessary for allthose who are involved in managing andorganizing information institutions and organi-zations ·like publishing houses, libraries, anddocumentation and information centres tounderstand certain characteristics of informa-tion. Understanding of these characteristics ofinformation like patterns of production, com-munication, dissemination, publicity and utili-zation is important if information has to beutilized optimally.

Bibliometric studies have been found veryuseful in assisting us to understand some of thecharacteristics of information as mentionedabove. The results of investigations of thesestudies have been found to be very useful to

34

DAVENDRA K. GUPTADepartment of Library, Archivaland Information StudiesUniversity of IbadanIbadanNigeria.

solve some of the practical problems of libra-ries and information centres. These studies havebeen proved to be very useful in planning thebudget for resource development, organizingeffective services, exploiting all possible infor-mation resources for the benefit of users and toaccelerate their activities. Numerous biblio-metric studies have been reported in the litera-ture of library and information science; butmost of them have focussed their attentionon the literatures of more or less stable fieldsof science and technology and social sciences,which are universal in scope. Recently, only afew studies have been conducted on the subjectof regional nature [1,2,7] and a very few onrelatively new areas of research [6,9]. Theresults of investigations reported for newlyemerging disciplines in science and technologyhave been found very fascinating and promptedme to study the characteristics and dynamics ofgrowth and development of another emergingresearch speciality in medical sciences - AIDS,hence this study.

AIDS

The epidemic of AIDS has been spreading likea wild fire all over the world during the lastfive years. A total of 31,646 cases of AIDSwere 'reported by the World Health Organiza-tion (WHO) from 74 countries. The largestnumber of the reported cases - 2.7,166 from 33countries, consisting of 86 per cent of thetotal - is from the Americans. Europe hasreported 3,127 cases from 22 countries, Africa1,008 cases from 10 countries, Oceania, 293cases from Australia and New Zealand, andAsia 52 cases from seven countries. TheseWHO figures were as of September, H~86 [11].But another recent report of WHO has givenupdated figures. Health authorities have notified34,448 cases of AIDS around the world as on14 November, 1986. WHO report mentionedthat 77 countries had notified AIDS cases while

Ann Lib Sci Doc

CITATION ANALYSIS OF AIDS LITERATURE

24 had reported that they found none; and that77 per cent of notified cases were in the UnitedStates [12]. This indicates a rising trend (evenwithin a period of two months) of the deadlydisease and its spread to countries hithertofree of it.

The· first AIDS case was documented inmid-1981 in the United States. AIDS is presentin many countries and in many populationgroups, .and will continue to spread unless effi-cient control programmes are established formassive public education and research. Statisti-cal dimensions and projections are alarming. Areport from the US Institute of Medicine "Con-fronting AIDS; Directions for Public Health,Health Care and Research" projects a cumu-lative total of more than 179,000 deaths fromAIDS in the United States in the end of 1991,with 54,000 of those deaths occuring in thatyear alone. It has been named the number onepriority of the United States Public HealthServices.

Acquired Immune Deficiency Syndrome(AIDS) is a serious condition characterised by adefect in natural immunity against disease. TheCentre for Disease Control (CDC) defines" AIDSas a disease, at least moderately predictive ofa defect in cell mediated immunity, occuringin a person with no know cause for diminishedresistance to that disease. AIDS is a disordercharacterised by T-Cell immunodeficiency, re-current and chronic infection with a variety ofsuch diseases as Kaposi's Sarcoma and Oppor-tunistic infection" [4]. In fact, as the nameimplies; AIDS is not a disease but a syndrome.Other details about AIDS like causes, symptoms,geographical distribution of cases, groups affect-ed, diseases. affecting AIDS patients and treat-ment etc. have been briefly described in anotherreport [8].

METHODOLOGY

Numerous accounts about AIDS have appearedin scientific literature and popular media sincethe first case was reported in June, 1981. Acollection of publications in a subject providesa good data base to extract enough quantitativedata to conduct a serious kind of structuraland bibliometric analysis. Therefore, a database was created by collecting information onall published scientific documents on AIDSduring the period 1981-1984. A bibliographywas compiled by Oji [10] under the personal

Vol 35 No 1 March 1988

guidance and supervision of the author, scanningall the primary and secondary sources of infor-mation available in the medical library of theUniversity of Ibadan (Nigeria). It was discoveredlater that the bibliography was not comprehen-sive for the year 1984 due to the fact thatmost recent materials did not reach in time(normal delay in obtaining the recent materialsin developing countries and Nigeria is no excep-tion for this). This obviously gave rise to alittle limitation. However, the data availablewith us was adequate to undertake bibliome-tric analysis.

Each item of the bibliography was consider-ed and relevant data on contributors, subjectareas, form of literature (books, reports, jour-nal articles, etc.), date of publication, etc. werenoted. Various permutations and combinationsof these parameters were carried out to studythe mutual relationships of these parametersand other characteristics of AIDS literature.

RESULTS AND DISCUSSIONS

Growth of the literature

The bibliography covered 316 items. Thesecomprised 273 journal items (86.4%) 19 books,9 conference papers, 9 audio-visual materials,4 reports and two chapters in books. Most ofthe non-serial items were published in 1984.A book published in 1981 is not directly relatedwith the literature of AIDS. The books andaudio-visual materials covered in this biblio-graphy are mostly of general nature. A closelook on the titles revealed that most of thebooks and A-V materials were created for pub-lic education and information describing generalaspects of AIDS. Only one conference washeld during this period - European Group forImmunodeficiencies Meeting at Chateau de'Fillerval (France) where all the nine paperslisted in the bibliography were presented.Since journal literature constitutes the majorproportion of AIDS literature, it has beendiscussed separately in another section.

From the data presented in Table 1, itappears that only 12 items were published in1981 - mostly as journal items. This numberincreased to 37 in 1982 registering an increaseof about 300%. This number increased dramati-cally and exponentially to 191 in 1983 (anincrease of 516% over the previous year). Thenumber of 1984 is not the correct reflection of

35

GUPTA

Table 1

Growth of AIDS literature

Forms of publications

Year Chapters Books Reports Conference A.V. Journal Total Percent- Cumulativein books papers items items age percentage

1981 0 1 0 0 0 11 12 . 3.8 3.8

1982 0 0 0 0 0 37 37 11.7 15.5

1983 0 8 3 0 3 177 191 60.4 75.9

1984 2 10 1 9 6 48 76 24.1 100.0

Total 2 19 4 9 9 273 316 100.0 100.0

Percent-age 0.6 6.0 1.3 2.9 2.8 86.4 100.0

the output as the entire literature did not reachthe Library, when the study was conducted.The number of publications for 1984, otherwiseshould have been much more than given in thebibliography. According to the trend, thisnumber should have been around 400 or so.This limited data, however, shows very clearlythe exponential growth of research output ofAIDS. This rate of growth is not surprising andconforms to the findings of two earlier studiesof similar nature [6,9].

Subject dispersion and Authorship trend

Entries in the bibliography were arranged al-phabetically according to the Medical SubjectHeadings (MeSH) of the National Library ofMedicine. In all, 16 subject headings wereidentified .to group all the items. Under eachsubject heading, authorship trends were alsostudied. The details of this analysis are presentedin Table 2.

Among all the subject headings, three promi-nent ones can be easily recognised. The firstone, Complications: the largest number ofpublications are listed under this subject heading- 70 (about 22%). The research activity in thearea is highly collaborative in nature, havingonly 9 single-authored items, while all otheritems are either corporate items or contributedby more than one author. The maximum num-

36

ber of authors i.e. eleven, contributing a pub-lication is listed under this category. Thesecond heading - Etiology has 41 items and outof these 16 items (39%) are contributed bysingle authors and this is the largest number ofsingle-authored items among all the categories;but it does not mean that this area is less colla-borative in research efforts. The third heading,Occurrence has a total of 38 items listed underthis. In fact, all the subject areas have beendeveloping as collaborative research efforts.This is not surprising for a speciality like AIDS,one of the areas of medical sciences which hasdrawn the world attention -- that most of theresearch is being carried out as team work inlaboratories, research institutions and hospitals.Only one area - Economics a social sciencerelated aspect has all the 3 publications assingle-authored items.

Journal Literature

Journal literature constitutes the major portionof the literature on AIDS, having a total of272 items out of 316, accounting for 86% ofthe total literature. Journal items were furtheranalysed to understand the kinds of itemsappearing in journals. 22 journals carried all thejournals items. Among the 272 journal items,journal articles were found to be 173 (64%);other items appeared as letters to editors - 83(30.5%); meetings - 8, editorial, and review

Ann Lib Sci Doc

CITATION ANALYSIS OF AIDS LITERATURE

Table 2

Authorship Trend and Subject Categories: AIDS literature, 1981-84

Subject Number of contributions made by number of authors Totalcategory

1 2 3 4 5 6 7 8 9 10 11 C._-----------------

1 2 2 1 5

2 9 9 10 9 4 6 4 4 2 0 1 12 703 3 4 7 5 1 2 22

4 1 1 1 3

5 3 3

6 16 7 3 1 3 2 1 1 1 6 41

7 1 2 1 1 5

8 2 2 3 0 3 3 0 0 1 14

9 5 0 1 3 4 2 3 1 19

10 8 3 6 5 0 1 3 1 2 9 38

11 1 0 2 3 0 1 2 9

12 5 1 2 3 11

13 1 1 1 1 4

14 1 5 1 1 1 1 1 11

15 1 1 2

16 19 14 2 1 1 2 1 19 59

Total16 78 49 37 28 18 20 14 8 2 6 1 55 316

Percent-age 24.7 15.5 11.7 8.9 5.7 6.3 4.4 2.5 0.6 1.9 0.3 17.4 100

Subject categories1. Blood 9. Microbiology2. Complications 10. Occurrence3. Diagnosis 11. Pathology4. Drug Therapy 12. Prevention and control5. Economics 13. Therapy6. Etiology 14. Transmission7. Familial & Genetics 15. Veterinary8. Immunology 16. General

Vol35 No 1 March1988 37

articles, notes and other items. Data on theanalysis are presented in Table 3. Journal itemsare shown against the respective journals inwhich they appeared and journals are arrangedaccording to the rank with total number ofpublications in each journal (this rank appearsin the rank column - a). Ranks band c werealso determined according to the total numberof journal articles and total number of lettersto the editors. Only 20 journals published allthe journal articles and only 8 journals publishedletters to the editors. Upto rank 7, almost allthe journals are common with different ranksat a, b and c. Lancet, New England Journal ofMedicine and Annals of International Medicineare the three top journals according to the ranka, band c. Incidentally, Nature and Science haveappeared in this list also. Lancet publishedlargest number of letters to the editors (38letters out of 83; in all 53 items were publishedin Lancet). Normally, letters to the editors arepublished in the professional journal mostpopular to the scientists. This has been verywell-reflected in this data. Three editorialspublished in the three most reputed journalsi.e. Lancet, New England Journal of Medicine,and British Medical Journal indicate the im-portance of this speciality at this early stageof development. All the editorials were publish-ed in 1983 just 2 years after the publication ofthe first report in 1981. One review article pub-lished in 1983 is little ahead of time and reflectsthe importance which the scientific communityis attaching to this subject.

Of the 22 journals, 10 journals published 10items or more, and seven journals publishedonly one item each. The first three journalstaken together published 126 items (46.2%).All the journals ranked upto 10, published238 items (about 87%). One could observethat most of the research output of this emer-ging speciality is confined to a very small num-ber of journals and scattering of informationis limited. Most of the ranked journals in thislist are well-known to the scientific communityand are subscribed by most medical libraries.

PRODUCTIVITY PATTERNS

Standard biblometric laws - Bradford's law forjournals' productivity patterns and Lotka's lawfor authors' productivity patterns were usedto test their applicability to the productivitydistribution data of this study.

38

GUPTA

Journals' productivity patterns

By half way through 1984, 173 items appearedas journal articles and 272 as journal items in22 journals. Two Bradford-Zipf bibliographswere drawn from the data. Bibliograph A wasdrawn using the data relating to all the journalitems and bibliograph B was obtained whenonly journal articles were considered. As dis-cussed before, the publications on AIDS areconcentrated only in a few journals and scat-tering of publications is almost negligible, itwas not logical to apply Bradford-Zipf typeof law. From the data and bibliographs, valuesfor "b", "s" and other parameters as describedby Brookes [3] were determined. The valuesof these parameters were found to be verymuch inconsistent and in some cases it wasnot possible to determine certain values. This,in fact, confirmed that the Bradford's law wasnot applicable in this case and perhaps, it isnot possible to apply the law to the productivitydistributions of journals in the case of an emer-ging research speciality, especially in the earlystage of its development. When plotted R (n)(cumulative total of papers) against log (naturallogarithm of the total of productive ~ournals),it was found that the data revealed elongatedS-shaped curves (Fig 1). Part one of the curve,the initial concave portion, represents the higherdensity of the nuclear zone; part two, the linearportion of the curve; and part three, often calledthe Groos droop, shows the deviation from thenormal and Brookes thought there were omis-sions from the relevant literature or to say thatbibliography on the subject was not complete[3,5]. However, a few observations can be madefrom the analysis of data and looking at thebibliographs in Figure 1: (1) that the two biblio-graphs are identical in shape and appearance,concluding that all the items appearing injournals behave in the same way in their distri-bution; (2) that the negative Groos Droop inboth the bibliographs confirms that the biblio-graphy is incomplete in its coverage for thisperiod; and (3) that the five journals form thecore in both the cases, contributing a total of170 journal items (about 62%) and 94 journalarticles (about 54%) respectively.

Productivity patterns of authors

A total of 671 authors have made contributionsto a total of 261 items. This means that on an

Ann Lib Sci Doc

Table 3

<!Rank List of Journals,' AIDS Literature J.B8.1-84

£ ----------------------------------------------------------------------------------~ Rank Name of the Journal Type of Materials Total Percentage Cumulative Cumulative(Jl ----------- --------------------------- Total PercentageZ a b c A L M E R N 00 ---------------------------------------------------------------------------------~

f 1 6 1 Lancet 13 38 1 1 53 19.5 53 19.5::r 2 3 2 New England J Med 19 22 1 42 15.4 95 34.9~~ 3 2 3 Ann internal Med 21 10 31 11.4 126 46.30000 4 1 - Morbidity Mor wkly 25 25 9.2 161 56.5

5 5 6 J Amer med Ass Rep 14 2 3 19 7.0 170 62.56 7 4 Nature 12 5 17 6.3 187 68.8 <.":)....7 4 - Federation Proc 15 1 16 5.9 203 74.7 ~0-3....8 9 - Clin Res 8 7 16 5.5 218 80.2 0

Z9 8 - AIDS Res 10 10 3.7 228 83.9

~10 10 6 J infections Dis 7 3 10 3.7 238 87.6 >e-11 11 7 Br med J 6 2 1 9 3.3 247 90.9 0<

en12i 11 Publ Hlth Rep 6 6 2.2 253 93.1 ....- 00

013 13 - AmerJ Med 4 4 1.5 267 94.6 ~13 13 Science 4 4 1.5 261 96.1 >- ....

I:'13 13 - WHO Rep 4 4 1.5 266 97.6 en

e-16 16 - Ann Allergy 1 1 6.4 266 98.0 ::3

to:!16 - - Arch internal Med 0 1 1 0.4 267 98..( ::a

~16 16 - Libert 1 1 0.4 268 98.8~16 16 - MedJ Aust 1 1 0.4 269 99.2 to:!

16 16 - Pediatrics 1 1 0.4 270 99.616 - 8 Radiology 0 1 1 0.4 271 100.616 16 - West Indian med J 1 1 0.4 272 100.0

22 20 8 TOTAL 173 83 8 3 1 1 3 272 100.0 272 100.0

PERCENTAGE 63.6 30.5 2.9 1.1 0.4 0.4 1.1 100.0

A '" Article; L '" Letter; M '"'Meetings; E - Editorial; R - Review; N" Notee; 0 '" Othell.CIJ a '" Rank according to the total number of publications\0 b '"' Rank according to the number of published articles

c '" Rank according to the number of letters to editors

~-In---;~~iUPPER ,NoNLINEAR I

• REGION &.l "';~I

: _,,1. ~I-G-R'~-O-S-1If------- I --------~ , DR•.•..•..~ __ II --~~ ,," i l .OOP,

I ",.,1LINEAR I L

REGION

NUCLEUS REGION

-- - - - - - - - - - - - - + --- - - ---,-I..

~ 250-~W•~ 200

II)s::0...•~u 150 -...•r-4.0-::3c:lt

aI~~ 100

50

GUPTA

,: 4ft,/ I GROOS

, I

,/ • ~OQ'- -- - -'..•.- +r > - ..••••-,-_ •.•• ·-1 L

I " _,I /.~_ •.•..__AII~

•lI

tjIII

IIII

IIIIIII I

I I

"AI Bibliograph for the data when allthe journal items counted

BI Bibliograph when only journalarticles counted

5 0loge(n) ie. loge L;J

Fig.i. Bradford-Zipf Bibliograph for AIDS journal titerature

average 2.5 authors contributed one item.Distribution patterns of authors were tested byapplying Lotka's law. Two different files werecreated - one for first authors only and theother was for the contributions of all theauthors (in this case, each author was given fullauthorship, if his name has appeared in theby-line of any title).

File 1: (only first authors counted)

The results of analysis are shown in Table 4.Values for a, Sand k were determined usingthe observed data and were compared with thetheoretical values. K-S statistical test was used totest the applicability of generalised Lotka's law.The distribution was found fitting to Lotka's

40

law with the value a = 3.4 at 0.01 level ofsignificance.

File 2: (for all authors counted)

The results of this analysis are presented inTable 5. Values for a, S, and k were determinedfrom the observed data -and were used to findtheoretical values. Observed and theoreticaldata were compared and tested with 'K-S statis-tical test for the validity of Lotka's law for thedistribution pattern of this data. The data werefound fitting Lotka's distribution with a = 3.2at 0.01 level of significance.

The maximum difference in observed andtheoretical values in both the cases were foundto be very negligible. The value of a = 3.4 and

Ann Lib Sci Doc

CITATION ANALYSIS OF AIDS LITERATURE

Table 4AIDS Literature, 1981-84: Proportion of Authors

(only first authors counted)a = 3.4S = 1.1384k = 0.784

Number ofcontributions

Observed S~(X) Expected [F o(X)-Sn(X»=D

1 0.9087 0.9087 0.8784 0.8784 0.0303

2 0.0696 0.9783 0.0832 0.9616 0.0167

3 0.0043 0.9826 0.0210 0.9826 0.0000

4 0.0130 0.9956 0.0079 0.9905 0.0051

5 0.0043 1.0000 0.0037 0.9942 0.0058

n = 230

D = [Fo(X) - Sn(X)) = 0.0303

At 0.01 level of significance, K-S statistics =1.63 1.63

15.17230

= 0.1075

Here, D < 0.1075, therefore, these data fit Lotka's law with a 3.4

Ta1Jle 5AIDS Literature, 1981-84, Proportion of Authors

(All the personal authors counted)

a = 3.2S = 1.1665k = 0.8573

Number ofcontributions

Observed Expected [F o(X)-Sn(X»= D

1 0.8614 0.8614 0.8573 0.8573 0.0041

2 0.0939 0.0939 0.0958 0.9506 0.0047

3 0_0224 0.9777 0.0255 0_9761 0.0016

4 0.0119 0.9896 0_0102 0.9863 0.0033

5 0.0045 0.9941 0.0050 0.9913 0.0028

6 0.0045 0.9986 0.0028 0.9941 0.0046

7 0.0000 0.9986 0.0017 0.9958 0.0028

8 0.0000 0.9986 0.0011 0.9969 0.0017

9 0.0015 1.0000 0.0008 0.9977 0.0023n = 671

D = [F o(X) - Sn(X)) = 0.00471.63 1.63

At 0.01 level of significance, K-s statistics =25.9671

=0.O62~Here, 0 < 0.0629, therefore, these data fit Lotka's law with Q = 3.2

Vol 35 No 1 March 1988 41

3.2 for two files are quite close to each otherand in any way more than 3. This confirms that,in the case of an emerging research speciality,a large number of authors is involved whocontribute only one paper each and hence thehigher value of this kind of author productivitypatterns. The value of Q is likely to decreasewith increasing stability of research specialityin future.

CONCLUSIONS

Some broad conclusions can be drawn from theanalysis and results of this study:

In 4 years time, since the first publicationappeared on AIDS in 1981, this newlyemerging research area has gathered consi-derable interest and. the growth of literaturehas been found to be dramatically exponen-tial after 1982. This trend is likely to conti-nue at least for a few years in future.

As expected, journal items constitute themajor portion of AIDS literature. A largenumber of letters to editors, editorialsand even one review article on the subject,indicate the active participation of scientificcommunity.

Sixteen different subject headings wereidentified, but the three areas namelycomplications, etiology and occurrencedominated and constituted the majorportion of AIDS literature.

Most of AIDS literature is confined to avery few professional journals and scat-tering is almost negligible. Bradford's lawdoes not apply to the journal productivitypattern but a core of five journals has beenidentified. At least one journal - AIDSResearch has been founded in late 1983in response to the increasing need for ajournal which concentrates on AIDS re-search. (Already 10 articles have beenpublished in the journal during this period).

A large number of authors have contributedto AIDS research. The data on authorproductivity pattern has fitted very wellwith the characteristic exponent value~ 3 to generalised Lotka's type distribution.The AIDS research speciality is progressing,perhaps, at the rate, at which the epidemicof the disease is spreading. Dynamics of

42

GUPTA

growth and other characteristics of thisfast growing research speciality need to beinvestigated further after a few years tounderstand the nature of the subject and itsinformation sources in a better way.

REFERENCES:

1. Aiyepeku W 0: Productivity of geographicalauthors: A case study from Nigeria. Journal ofDocumentation 1976, 32, 105·112 .

2. Alabi G A: Bradford's Law and the Nigerianentomological literature. Insect Sci Application1984,5,7-11.

3. Brookes B C: Derivation and application of theBradford-Zipf distrbution. Journal of Documen-tation 1968,24,247-265.

4. CDC: Update on aquired immune deficiencysyndrome (AIDS) - United States. MMWR 1982,31, 508, 513.

5. Groos 0 V: Bradford's Law and the Keenan-Atherton data. American Documentation 1967,18, 46.

6. Gupta D K: Structural analysis of early literatureof computer-based storage and retrieval of geo-science information. IASLIC Bulletin 1984, 29,1-10.

7. : Lotka's law and productivity patternsof entomological research in Nigeria for theperiod, 1900-1973. Scientometrics 1986, 12(1-2),33-46.

8. : AIDS: Development and growth of anewl y emerging research speci ali ty in medicalsciences. (in press).

9. Lawson J, Kostrewski B and Oppenheim C A A:A bibliometric study on a new subject field:Energy analysis. Scien tometrics 1980.:l. 227 -237.

10. Oji K A: AIDS, 1981-1984: A bibliography 1986;31p. Ibadan, University of Ibadan, Department ofLibrary Studies (Unpublished Report).

11. WHO: 31,646 cases of AIDS reported. Times ofIndia 24 October, 1986; 28 (News).

12. WHO : AIDS. Sunday Times, 23 November,1986; 2g (News).

Ann Lib Sci Doc