Family Practice Advance Access originally published online on August 26, 2005
Family Practice 2005 22(5):570-575; doi:10.1093/fampra/cmi085
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
PubMed perspective of family medicine research: where does it stand?
a Faculty of Medicine, University of Kelaniya, Sri Lanka, b Computer Centre, Faculty of Medicine, University of Kelaniya Sri Lanka and c Present address: School of Rural Health P.O. Box. 1043 11, Moran Drive Dubbo NSW 2830 Australia
Correspondence to K. Mendis; Email: kmendis{at}med.usyd.edu.au
Received 12 January 2005; Accepted 2 August 2005.
Mendis K and Solangaarachchi I. PubMed perspective of family medicine research: where does it stand? Family Practice 2005; 22: 570–575.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Objective. The aim of this study was to obtain a view of family medicine research by analyzing PubMed citations from 1960–2003.
Method. Family practice (FP) citations in PubMed from 1960 to 2003 were downloaded in MEDLINE format. This was written into relation database using ‘PubMed Grabber/Analyzer’ software developed at University of Kelaniya, Sri Lanka. Search Query Language (SQL) and online PubMed queries were used for further analysis.
Results. There were 50288 FP citations from 80 countries. Of these, 33712 (67%) citations were from 15 FP journals. United Kingdom (18760), United States (13584), Australia (3262), Canada (1848), Germany-west (1340) were the five countries which had the most citations and 22 countries had less than 5 citations. Van Weel C (118), Geyman JP (116), Olesen F (87), Jones R (83) and Knottnerus JA (82) were numerically, the top five authors. Only 921 authors had more than 10 citations and the vast majority of authors had only one citation. Letters (5121), review (2715), editorial (2259), randomized controlled trials-RCT (1585) and Meta-analysis (44) were the top publication types. 40 citations found under ‘qualitative research’.
Discussion. The relatively few PubMed FP citations (50288) are by a small number of academics in developed countries. Citations showed an upsurge from the mid 1980s to the late 1990's but reached a plateau in the new millennium.
Compared to PubMed citations from 1960–2003 in other specialties such as 2737655 for public health, 1151194 for cardiology & cardiovascular diseases and 318538 for medical informatics, the 50288 FP citations were paltry. Paucity of RCT (1585) and meta-analysis (44) was noted. The low ‘qualitative research’ citations (44) could have been due to the late introduction of the MeSH concept in 2003.
Conclusions. Priority should be given to increase FP research and also to ensure the indexing of FP journals that are not currently indexed in PubMed. Efforts to increase citations in Medline may not give the desired results because of low priority given primary care specialties such as family medicine in the USA. Alternative solution of a separate bibliographic database for FP similar to PsycInfo may be too costly.
Keywords. Family Medicine, Family Practice, General Practice, PubMed, Medical Informatics.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Family physicians are in increasing numbers accessing electronic databases for their literature searches and also to update themselves. Citations in the electronic databases will ensure a wider coverage of the published literature. Where citations in the electronic databases are concerned there is a perception that family medicine receives Cinderella treatment. If this is true it will no doubt retard the steady development of this discipline, which though it is as old as Hippocrates is paradoxically a young discipline. Family Medicine was recognized in the United States of America by the American Medical Association as the 20th discipline only in 1969. In the UK, the first chair in general practice was established in as late as 1963. The World Organization of Family Doctors (Wonca) was founded in Melbourne in 1972 and today is the largest organization of family physicians/GPs having members from 79 countries representing about 80% of the world's population.
Practice of medicine is defined as "... an art, practiced against a background of incomplete scientific knowledge about the nature of the disease process".1 Nevertheless the current medical practice has been inclined towards the ‘science’ component.2 Despite reservations from the conservatives evidenced-based medicine (EBM) has been emerging as the new paradigm in the practice of medicine. The majority of family physicians are advocates of the ‘art’ of medical practice. Many family physicians regard research as a "specialist endeavour or a task necessary for the few family physicians who want to become academics".3
It has been shown that to bridge the research–practice gap, family practice research must expand its knowledgebase in different ways that maybe unique to family practice.4 To address the issue of the importance of research in family medicine, Wonca recently organized the Wonca Kingston conference.5
Medline which is the largest medical electronic bibliographic database in the public domain was produced by the National Library of Medicine (NLM) in USA. It has been identified with jazz music as "one of two greatest gifts that America has given to the rest of the world".6 Medline contains more than 12 million citations from 4800 journal articles in life sciences with a concentration on biomedicine.7 More than 40 full-time indexers, including content as well as information science experts, add more than 500 000 biomedical article citations to Medline annually. In 1997 worldwide free access to Medline through the internet was initiated by the NLM and was named PubMed (http://www.pubmed.gov). Medline, the largest public domain electronic knowledgebase in biomedicine, is the starting point of many systematic reviews, meta-analyses and practice-guidelines that brings biomedical research findings to the point of care.8
We analysed family practice citations in PubMed from 1960 to 2003 to obtain an in-depth view of family medicine research. The term family medicine (FM) will be used throughout this article to denote the discipline general practice/family practice.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Medline is available as an on-line resource in the Internet, free of charge at PubMed (http://www.pubmed.gov). The search strategy for family medicine citations was to increase the sensitivity. In this there were three basic queries.
The first query was aimed at obtaining all citations which had ‘family practice’ OR ‘general practice’ OR ‘family medicine’ as a ‘MeSH’ OR ‘Text’ word from 1960/01/01 to 2003/12/31. The detail query is set out below:
(("family practice"[TIAB] NOT Medline[SB]) OR "family practice"[MeSH Terms] OR general practice[Text Word] OR family medicine[Text Word]) AND ("1960/01/01"[PDAT]: "2003/12/31"[PDAT])
The second query obtained all citations, which had ‘general practice, dental’ as a MeSH word from 1960/01/01 to 2003/12/31. The detail query is set out below:
"General Practice, Dental"[MeSH] AND ("1960/01/01"[PDAT]: "2003/12/31"[PDAT])
Finally the third query combined the two searches by using the Boolean term ‘NOT’ to obtain only the family practice/general practice citations.
We also looked at related MeSH words that would overlap with the discipline of GP/FP. ‘Primary care’ or ‘Primary health care’ when searched in PubMed gave 52 450 hits for the time period 1960–2003. We searched for the words ‘general practice’ OR ‘family practice’ OR ‘family medicine’ in the subset of 52 450 (Text word search), and only obtained 7827 hits. When we combined the results of our third query and ‘primary health care’ we brought in about 44 623 citations that did not have the words family practice/general practice/ family medicine into our analysis. Therefore we did not use the ‘primary health care’ search term in our search query.
We downloaded the data in Medline citation format, which is a text file that can be read using any word processor such as Microsoft Word or Notepad. A Medline citation consists of many fields of information such as title, author and source which is similar to the structured abstract of a journal article. Other structured information such as MeSH words, country, author and affiliation is included by the NLM staff at the time of indexing.
Using a software application (PubMed Grabber/Analyzer-PGA)9 developed at the Computer Centre, Faculty of Medicine Ragama, Sri Lanka, the important fields in the Medline text file were written to a relational database management program in Microsoft Access 2000. The citation fields that we captured were: unique identifier (UI), publication ID (PMID), title (TI), authors (AU), affiliation (AD), source (SO), date of publication (DP), publication type (PT), country of origin (CY), Medical Subject Headings (MH), Abstract (AB). The analysis of this database was done using PGA.
We used on-line PubMed queries to ascertain the number of citations for other disciplines such as public health, cardiology and cardiovascular diseases, medical informatics for comparison. These queries were formulated so that the specificity was maximal on each of the disciplines. The PubMed journal database was also searched on-line to obtain the details of the journals published.
The results that we obtained using PGA were crosschecked with direct on-line PubMed queries wherever it was possible to do so, for example (a) author, country frequencies (b) the publication types for family practice. A paper on the methodology and application software of PubMed Grabber/Analyzer has been published elsewhere.9
All the web sources and PubMed queries were accessed in November 2004.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
There were 50 288 family medicine (FM) citations from 80 countries for the period 1960–2003. Over the past four decades the growth of FM citations has been slow (Table 1). There has been a relatively rapid growth of citations from 1987–1998. The number of citations for the period 1996–1998 was 7058 and for 1999–2001 was 7329, an increase of only 271. In the last completed three year period of 1999–2001 it has almost reached a plateau. The apparent reduction in the period 2002–2004 is because the data was tabulated only for two years up to the end of 2003.
|
PubMed had indexed a total of 23 FP journals. We have listed the 15 journals that had more than ten citations during the period 1960–2003 (Table 2). Some of the journals were discontinued by 1960. The American Family Physician (7932), Journal of Family Practice (6296), Australian Family Physician (6077), British Journal of General Practice (3482) and Canadian Family Physician (2493) were the five journals that had the largest number of citations. A total of 33 712 (67%) citations were from these 15 family practice journals. For the non-FP journals the British Medical Journal had the highest FP citations (1024).
|
Five countries with the highest number of citations were England (18 760), the United States (13 584), Australia (3262), Canada (1848) and Germany, west (1340) (Table 3). As many as 22 countries had only five or less citations.
|
The five authors with the highest number of articles were van Weel C (118), Geyman JP (116), Olesen F (87), Jones R (83) and Knottnerus JA (82) (Table 4). 921 authors had more than 10 citations but that the vast majority of authors had only one citation.
|
The citations categorized according to the PubMed ‘publication types’ that is available on the drop-down menu in the ‘Limits’ are set out in Table 5. The letters (5121), review (2715), editorial (2259) were the top three publication types. There were 1585 randomized controlled trials (RCTs), 2248 clinical trials and 44 meta-analyses.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
This study shows that the relatively few PubMed family medicine citations are by a small number of academics in developed countries. Compared to PubMed citation from 1960–2003 in other specialties such as 2 737 655 for public health, 1 151 194 for cardiology and cardiovascular diseases and 318 538 for medical informatics, the 50 288 FP citations were paltry. Citations in subspecialties such as cardiology trigger a huge number because of their disease component, for e.g. cardiovascular disease component comprises 1 146 977 citations and Cardiology only 6875. Family Medicine does not have such a large disease component.
When some disciplines such as public health show an almost exponential increase in citations, family practice citations in the new millennium has reached a plateau (Table 1). However there has been a reported increase in family practice citations in Australia10 and The Netherlands.11
There were 373 581 clinical trial, 183 778 RCT and 8592 meta analysis citations in PubMed for the period 1960–2003. The comparative family practice citations shown in Table 5 were: clinical trials 2248 (0.6%), randomized clinical trials 1595 (0.9%) and meta analysis 44 (0.5%). Out of the 1585 RCTs in family medicine, 879 (55.5%) were from UK, 295 (18.6%) from USA and 55 (3.5) from Norway. The other 25 countries together had only 356 (22.5%) RCT citations. Though on-line PubMed searching may not be the most appropriate method of identifying RCT citations, it is nevertheless the most cost effective way of searching the literature. However hand searching may be the most accurate method.12
There is little distinction made between ‘non-research’ and ‘research’/‘papers’ in our analysis, for example most of the articles in Australian Family Physician will be categorized under ‘non-research’. In comparison to the Australian Family Physician, a larger number of articles in the British Journal of General Practice will come under the ‘research’/‘paper’ category. There is no way of distinguishing these two broad categories using PubMed queries. However some may argue that even to write a ‘non-research’ or ‘review’ type of article one has to research the literature.
Qualitative research is more in conformity with the methods and processes of family practice. Surprisingly out of the total of 1167 citations of qualitative research in PubMed for the period 1960–2003, only 40 (3.4%) citations were family practice. One reason maybe because the MeSH concept ‘qualitative research’ was only introduced into PubMed as late as in the year 2003.
Reasons for the paucity of FP citations are many:
- Family practice research is static and not increasing. Compared with the secondary and tertiary care specialist, family physicians do little research.13 Wonca has already addressed this issue recently with the Kingston conference.5
- Family practice journals, especially the non-USA and journals in languages other than English do not receive adequate indexing in PubMed. Of the four FP journals indexed in or after the year 2000 three of them are from USA (Table 2).
- The MeSH/Text words ‘Family Practice’/‘General Practice’/‘Family Medicine’ may miss out on a number of research papers published under the MeSH/keywords ‘primary health care’ (PHC). However family practice/general practice is just one specialty that provides PHC. One way of addressing this issue in the future is to include ‘family practice’ as one essential MeSH word in the articles submitted to Medline indexed journals. This problem should be addressed by an international committee of researchers in FP/GP and recommendations made to the relevant authorities including National Library of Medicine—PubMed.
- Joint research studies with other disciplines and research done by family physicians with areas of special interest may not be shown in this type of analysis. This harms our current position and may be even worse. Family medicine efforts may lead to benefits for other disciplines; collaboration undermines rather than strengthens our position as no visible discourse in our discipline results from it!
In the foreseeable future the US health care will be demand oriented. As America's affluence increases there will be a hue and cry for more subspecialty services.14 These factors are already leading the US healthcare service to a primary care workforce crisis. Therefore it is unlikely that the sub-specialty biased PubMed will increase indexing family practice journals. It is debatable that the 6th recommendation of the Wonca Kingston conference15 which states "... a wider representation of family medicine research journals in databases like Index Medicus should be pursued" will have a major influence on the number of citations. Efforts by Wonca to increase indexing of family practice journals in PubMed since 2000 may have reached its summit with four new journals being indexed since 2000 (Table 2). Despite the efforts the total number of citations has reached a plateau in 2003 (Table 1).
A multi-pronged effort to increase and also utilize family practice research that has already being published is needed. One way is to incorporate the research already published in non-Medline indexed journals. In two family practice websites, Wonca global family doctor (http://www.globalfamilydoctor.com/) and Primary Care Internet guide (http://www.uib.no/isf/guide/journal.htm), we found 45 family practice journal links, with some allowing on-line full text access. With the discontinued ones, PubMed has only indexed 23 FP journals with the discontinued ones.
If prospects for indexing family practice journals are not forthcoming in PubMed, a logical step would be to setup an electronic bibliographic family practice database similar to PsycInfo (http://www.apa.org/psycinfo/). Psycinfo is an electronic bibliographic database for behavioural science and mental health with nearly two million records indexed from journals as far as 1887. This is a rapid way of utilizing the already available publications without having to wait for an increase family practice publications. The cost of setting up such a database and maintaining it would be prohibitive for a single organization to attempt. A regional solution has been suggested that we share the information and raise the profile of family medicine in the Asia Pacific Region of WONCA.16
This study shows that the top ten developed countries contributed 87% of family practice citations in PubMed. South Africa (323) and India (186) were the only two developing countries among the first twenty. There were only 54 citations from the African continent other than the South African citations. The FP research from developing countries has to be encouraged if appropriate guidelines regarding cost-effective primary care are to be implemented in these countries. The context of practice and patient preferences maybe different and the same guidelines may not be applicable to both developed and developing countries.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Funding: self.
Ethical approval: all information used was in the public domain. There was no patient intervention.
Conflicts of interest: none.
|
Acknowledgments |
|---|
The authors thank Dr Dennis J Aloysius for editing the article.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
1 Weatherall DJ, Warrell D. Oxford Textbook of Medicine. Oxford: Oxford University Press; 2003, 20.
2 Cochrane AL. Effectiveness and efficiency: random reflections on health services. London: Nuffield Provincial Hospitals Trust; 1972.
3 Bentzen N. Family medicine research: Implications for Wonca. Ann Fam Med 2004; 2: S45–S49.
4 De Maeseneer JM, van Driel ML, Green LA, van Weel C. The need for research in primary care. Lancet 2003; 362: 1314–1319.[CrossRef][Medline]
5 Rosser WW, van Weel C. Research in general/family practice is Essential for Improving Health Globally. Ann Fam Med 2004; 2: S2–S4.
6 Smith R, Chalmers I. Britain's gift: a "Medline" of synthesized evidence. Br Med J 2001; 323: 1437–1438.
7 National Library of Medicine. Facts Sheet. National Library of Medicine. http://www.nlm.nih.gov/pubs/factsheets/nlm.html (accessed on 19–11–2004).
8 Haynes R. Of studies, syntheses, synopses, and systems: the "4S" evolution of services for finding current best evidence. Evidence Based Med 2001; 6: 36–38.
9 Mendis K, Solagarchchi I, Weerabaddana C. PubMed Grabber/ Analyzer—a tool for off-line analysis of PubMed. J Korean Soc Med Informatics 2003; 9(Supp 2): s249–s244.
10 Ward A, Lopez D, Kamien M. General practice research in Australia, 1980–1999. Med J Aust 2000; 173: 608–611.[Medline]
11 Knottnerus JA. Research in general practice. Lancet 1996; 347: 1236–1238.[Medline]
12 Jadad AR, MqQay HJ. A high-yield strategy to identify randomized controlled trials for systematic reviews. Online J Current Clinical Trials 1993 Feb 27; Doc No 33.
13 Askew DA, Glasziou PP, Del Mar CB. Research output of Australian general practice: a comparison with medicine, surgery and public health. Med J Aust 2001; 175: 77–80.[Medline]
14 Phillips RL, Stratified B. Why does a primary care physicians workforce crisis matter? Am Fam Phys 2004; 70: 440, 442, 445–446.[Medline]
15 Van Weel C, Rosser W. Improving health globally: the necessity of family medicine research—a critical review of its implications and recommendations to build its capacity. Ann Fam Med 2004; 2: S5–S16.
16 Teng SL. Is an indexed database publications database for the Asia Pacific region feasible? Asia Pacific Fam Med 2002; 1: 33–36.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. J. Carek and A. G. Mainous III The State of Resident Research in Family Medicine: Small but Growing Ann. Fam. Med, January 1, 2008; 6(suppl_1): S2 - S4. [Full Text] [PDF]
|
|
E-letters:
Read all E-letters
- Cave PubMed
- Tzeng-Ji Chen
- Family Practice, 20 Sep 2005 [Full text]
- PubMed perspective of family medicine research
- Anders Håkansson, et al.
- Family Practice, 16 Nov 2005 [Full text]
- Håkansson & Kragstrup's comment
- Kumara Mendis
- Family Practice, 12 Dec 2005 [Full text]
- Re: Cave PubMed
- Kumara Mendis
- Family Practice, 19 Dec 2005 [Full text]
- Bibliography of primary care research in Malaysia
- Cheong Lieng Teng, et al.
- Family Practice, 30 Dec 2005 [Full text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||