Key-words: Bibliometric analysis - neurosurgery research - Pakistan
Introduction
The global distribution of healthcare-related research efforts is unequal between
high- and low-income countries.[[1 ]] This has been highlighted on multiple global forums since 1990. The World Health
Organization Consultative Expert Working Group's report on financing and coordination
of research and development related to diseases that mainly affect the world's poorest
people demonstrate ongoing disparity.[[1 ]] Global burden of disease data and publications from developing countries both suggest
that spectrum of diseases affecting developing countries is different from that of
developed countries.[[2 ]] Yet, clinical trials on diseases of relevance to high-income countries are seven-to-eight
times more often than those related to low- and middle-income countries.[[3 ]] As funding opportunities for developing countries decline over the years, it has
become imperative for developing countries to make special efforts.[[4 ]] Each health specialty in developing countries needs to go through a process of
rigorous self-assessment to identify and address areas of deficiency.
In 2010, the senior author of this study analyzed the overall research output of neurosurgical
community of Pakistan.[[5 ]] The results were presented at multiple forums and published in an article. Pakistani
neurosurgeons published an average of only ten PubMed indexed papers per year. There
were neither clinical trials nor systematic reviews or basic science research. This
generated much debate within the neurosurgical community, and a number of suggestions
were floated to encourage and promote nationwide neurosurgical research.
Over the past few years, the country has seen an increase in the number of neurosurgeons
with an apparently greater emphasis on clinical research. To determine if there has
been an increase in the number of peer-reviewed articles in neurosurgery from Pakistan,
we decided to do a repeat bibliometric analysis of the past 6 years. We also chose
one of the Pakistani neurosurgical departments on the basis of its relatively outstanding
research contribution to draw a comparison with neurosurgical departments around the
world, using h and i-10 indices.
Methodology
We conducted a bibliometric analysis of publicly available databases for all neurosurgery
publications from Pakistan. This study was exempted from Institutional Review Board
by the Ethics Review Committee of the Aga Khan University Hospital (AKUH). All indexed
peer-reviewed publications from January 2009 to December 2014 where at least one author
was affiliated with any of the neurosurgical departments in Pakistan and research
was conducted in Pakistan were selected. Both international and local scientific search
engines (including PubMed/Medline, Cochrane database, Embase, Google Scholar, PakMedinet)
were used in various combinations of keywords, including both text words and medical
subject headings (MeSH). Search for focused topics was performed by grouping MeSH
terms according to the associated neurosurgical subspecialty. Furthermore, names of
various neurosurgical departmental chairs along with their respective center and major
Pakistani cities were used separately as keywords. Senior neurosurgeons and unit heads
of major neurosurgical institutes were also contacted individually in order not to
miss publication. Locally published indexed journals, Journal of Pakistan Medical
Association, Journal of College of Physicians and Surgeons Pakistan, and Journal of
Ayub Medical College were manually searched for neurosurgery publications.
All publications that were not directly related to neurosurgery or in which data were
acquired from outside of Pakistan were excluded. Nonscientific correspondence such
as bibliographies, news items, and roll calls of reviewers was also excluded. The
resulting references were screened on the basis of abstracts and full texts where
necessary. This method has been described in detail previously.[[5 ]] Selected publications were evaluated for basic characteristics. Frequencies and
proportions were calculated for publication with respect to study design, specialty,
city, institute, and authors. All data were recorded using Statistical Package for
Social Sciences version 20 (SPSS IBM, Chicago, IL, USA).
The neurosurgery department with the most publications was then selected, and several
variables including i-10 index, h-index, number of citations, and publications per
faculty per year were calculated using Google Scholar. These numbers were then compared
with published reports of international neurosurgical centers.
Results
Follow-up bibliometric analysis
Neurosurgeons from Pakistan published 121 articles during the defined study period
(mean = 20.16 ± 5.2 papers/year). A relatively constant increase was noticed for the
last 6 years, i.e., 2009–2014. [[Figure 1 ]] identifies the number of publications during the study period along with publications
from previous years. Majority, i.e., 100 (82.4%) of these publications came from a
single city of the country.
Figure 1: Trend of neurosurgery publications per year
Neurosurgeons primarily authored 111 (91.7%) studies while the rest were collaborations
with other specialties such as neurology, internal medicine, and basic sciences. Departmental
affiliation of authors could not be identified in two publications. The mean number
of authors per publication was 4.2 ± 1.4. Sixty-two (62.8%) of the articles were published
in international journals, and 45 (37.2%) were published in national PubMed indexed
journals. These figures showed that Pakistani neurosurgeons were now more likely to
choose an international journal than the last audit.
We stratified articles according to the study design. Case series, case reports, letters
to the editor, and forum articles constituted the largest segment of these publications
(n = 58, 47.9%). Original articles (retrospective and prospective cohort studies)
constituted 47.1% (n = 52) of all publications. There were 5 review articles, 1 meta-analysis,
and one quasi-experimental study [[Table 1 ]]. This was similar to the trends of global neurosurgical publications.[[6 ]] In terms of specialty, neuro-oncology (n = 28, 23.1%), spine (n = 23, 19.0%), and
neurosurgical trauma (n = 19, 15.7%) had most number of publications [[Table 2 ]]. Eight articles addressed hydrocephalus; one of the studies was related to basic
sciences and only two studies were funded. When compared with global neurosurgical
publication trends, we found fewer publications on vascular and functional neurosurgery
and a comparatively larger proportion of publications on central nervous system infections
and hydrocephalus.[[6 ]]
Table 1: Number of articles according to the study design
Table 2: Publications with respect to subspecialties
Comparison of Pakistani neurosurgical center with international centers
When we further analyze the Pakistani neurosurgical department with the highest publication
rate, the AKUH in Karachi, we found that during the study period, the department contributed
80 of 121 (66.1%) publications coming from the country (rate of 16 per year). AKUH
is a private university hospital with ISO certification and JCIA accreditation, and
the neurosurgery department has six full-time faculty members. For the study period,
the department showed a publication/faculty/year rate of 2.2. Three primary authors
of the department contributed to 76 (62.8%) of all publications coming from the country.
For the study period, the mean h-index of the department is 10, mean i-10 index is
13, and the cumulative i-10 index is 78.
Here, we must mention that these comparisons in no way compare the value of the department
or the value of the research conducted there and are subject to the inherent limitations
of the various tools for measuring research outputs.
India
On comparing these statistics with neurosurgical centers in India, we found a similar
pattern of publication where majority of research is conducted in just a few centers.[[7 ]] The publication rate of AKUH (16 per year) was much higher than the publication/year/institute
rate of India – 3.28 ± 8.22. The Indian data have been analyzed for private and government
centers, separately. If AKUH is to be ranked on their tables, it would rank higher
than any other private center and would be one of the top 7 centers with respect to
total publications in 5 years, total number of citations, number of original papers
in the past 5 years, 5-year i-10 index, and 10-year h-index.
Saudi Arabia
The mean h-index of neurosurgical centers in Saudi Arabia is 5.04 (range 0–33),[[8 ]] and although an individual analysis of various center was not available for comparison,
AKU with a mean h-index of 10 (sum of h-indices 60) would certainly rank among the
top centers in Saudi Arabia.
Great Britain and Ireland
h-index is a function of time. The older institutes would have a higher h-index. Few
of the oldest neurosurgical centers are in Great Britain and Ireland. On comparison
of the mean h-index and sum of h-indices, AKU will be ranked easily in the top 10
neurosurgery centers in Great Britain and Ireland.[[9 ]]
United States of America
A significant proportion of global neurosurgical publications come from the USA.[[6 ]] However, the relative rates of publication of the USA have remained constant where
other countries have shown several hundredfold of improvement in their contributions.[[6 ]] Similarly, several countries, despite having fewer publications, tend to have more
impactful publications.[[6 ]] Nevertheless, the USA continues to be the global leaders in neurosurgical research.[[10 ]],[[11 ]] If AKUH was to be ranked on the tables of North American neurosurgical centers,
it would lie someplace between 60 and 70, on the basis of i-10[[5 ]] index, summed h index, total publications, and total citations. However, most North
American centers on the list cannot be compared to AKUH by virtue of their size, and
when AKUH is compared to centers of its own size (8 or less faculty members), it would
lie within the top 10 neurosurgical centers in the USA.
Discussion
Pakistan has a population of 190 million and is the sixth biggest country in the world.[[12 ]] The presence of less than 150 neurosurgeons in the country, gives it one of the
lowest surgeon to patient ratios in the world, higher only to a few African countries.[[13 ]],[[14 ]] This points to an extremely high patient load per surgeon. Not surprisingly, the
quality and quantity of neurosurgical and neuroscience research in Pakistan, considering
its size, has been below par.[[15 ]]
The study in 2010 which analyzed the number of neurosurgical publications over a 6-year
period identified only about 10 papers from Pakistan per year.[[5 ]] Results of the current study indicate a steady increase in the number of neurosurgery
publications from Pakistan. Compared to the previous study, the number of neurosurgery
publications has doubled in the last 6 years. Majority of the publications however
were still contributed by a few institutes only. Several of the institutes did not
have a single publication during the study period.
Our follow-up bibliometric analysis reveals a rising trend in research publication
with an annual growth rate of almost 17%. The proportion of publications in international
journals has increased by almost thrice the amount (from 22.6% to 62.8%) with a consequent
drop in local journal articles (77.4% to 37.2%). Such a finding indicates improvement
in literature coming from the country with publication in greater impact factor journals.
This improvement in quality is also implicated by the publication of higher evidence
studies such as meta-analysis and case–control studies that were seen during our study
period.[[16 ]],[[17 ]] Trauma, infections, and hydrocephalus are leading neurosurgical problems in developing
countries. These are not reflected as priority in the publication profile of the country.
Only 15.7% papers concerned trauma while 8 out of 121 articles addressed hydrocephalus.
According to King's analysis, 31 nations produce 98% of the volume of scientific citations
in the world and the rest of the world contribute to the remaining 2% citations.[[18 ]] In another study on biomedical research profile of nations, Pakistan was among
the countries with lowest research profiles (0.86 population per million population).[[19 ]] Neighboring country India had a much better research profile with 2.82 publications
per million population. Developed countries like Japan produced over 240 publications
per million population but several low-income African countries such as Kenya, Libya,
Gambia, and Gabon had higher number of publications per million population, than both
Pakistan and India.[[19 ]] This shows that although a better economic status of countries helps, it is possible
to generate research despite limited resources. Through this paper, we have shown
that although Pakistan may lag behind in overall research outputs, and neurosurgical
research outputs, one center in the same country is still able to produce reasonable
research, despite the limitations, proving that these limitations can be overcome.
Indeed, several developing countries have been shown to have several hundredfold increases
in neurosurgical research outputs over very short period of time.[[6 ]] Another prime example is the University of Toronto Neurosurgery Program, which
in a span of just 5 years, increased its research outputs from a time that it was
comparable to AKU, to now where it ranks even higher than the top program in the USA.[[11 ]] Salgar outlined three key reasons for the low research output in developing countries,
lack of scientific writing training, lack of budget for specialized editorial staff,
and lack of expert assistance, which we believe are very valid although our comparison
also suggests a fourth reason, which is the smaller number of faculty members within
the department.[[20 ]]
There are several important limitations of our study. Since we have restricted our
research to single specialty, it is not a true reflection of overall state of research
in Pakistan. A comparison with other specialties would have been useful. For our bibliometric
analysis, we only included studies that were indexed in Medline, just as it was done
for the last review, which excludes all studies published in nonindexed journals.
Second, spine surgery is also performed by spine orthopedic surgeons, and their papers
coming from Pakistan were excluded. For the institutional comparison part of the study,
we must admit that it is limited by the well-discussed, inherent limitations of the
tools available for comparison, i.e., h-index, i-10 index, etc.[[21 ]],[[22 ]]
Conclusion
Neurosurgery research in Pakistan has shown modest improvement in terms of quality
and quantity. There exists a large disparity between institutes in scientific research
output within the country. A homogeneous distribution of resources, strong will on
the part of clinicians, and collaboration between institutes can help improve the
scientific output from the country.
