Electronic Letters to:
|
|
E-letters: Electronic letters published:
-
![[Read eLetter]](/icons/misc/sectionDOWN.gif)
Bibliography of primary care research in Malaysia
- Cheong Lieng Teng, Ee Ming Khoo, Chirk Jenn Ng (30 December 2005)
-
Re: Cave PubMed
- Kumara Mendis (19 December 2005)
-
Håkansson & Kragstrup's comment
- Kumara Mendis (12 December 2005)
-
PubMed perspective of family medicine research
- Anders Håkansson, Jakob Kragstrup (16 November 2005)
-
Cave PubMed
- Tzeng-Ji Chen (20 September 2005)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Cheong Lieng Teng, Associate Professor International Medical University, Ee Ming Khoo, Chirk Jenn Ng
Send e-letter to journal:
|
We read "PubMed perspective of family medicine research"[1] and the related rapid responses with interest. We noted the discussion on differing definitions of "family medicine research" and what journals can be categorised as "family medicine journals". In this letter, we wish to touch on Mendis and Solangaarachchi's documentation of the paucity of family medicine research papers from developing countries that were indexed by PubMed. In response to the call by Liaw,[2] we have started to collate original primary care research conducted in Malaysia. Our bibliographic work attempts to retrieve publications of original research conducted within the Malaysian primary care settings, i.e. general practice clinics (private), health centres (publicly funded) and community (e.g. survey of the general public or school populations) between 1966-2004. Our methods included hand search (of Malaysian and selected non-Malaysian English language journals) and database search (PubMed and other local university databases) and followed by a selection of retrieved abstracts/full texts based on the consensus of three primary care researchers. After a two-year search, we identified only 602 original articles and 186 theses/dissertations satisfying our criteria of Malaysian primary care research publications. The retrieved original articles came from 73 journals (37 of these journal are indexed in PubMed, contributing 273 articles). In only 17 of these articles have the words "family practice" or "general practice" appeared in the titles/abstracts. Our experience points to the large body of Malaysian primary care research papers in the non-indexed publications. The keywords "family practice" or "general practice", used by Mendis and Solangaarachchi in their PubMed search, have very low sensitivity for the retrieval of primary care research publications in the Malaysian context. Our choice of "primary care settings" allow inclusion many more research papers that more closely match the needs of primary care researchers, especially in a developing country like ours where there is clear public-private division in the provision of primary care and family medicine is still striving to establish itself. Reference: 1. Mendis K,.Solangaarachchi I. PubMed perspective of family medicine research: where does it stand? Fam Pract 2005;22:570-5. 2. Liaw ST. Is an indexed database publications database for the Asia Pacific region feasible? Asia Pacific Family Medicine 2002;1:33-6. Teng CL, International Medical University (tengcl@gmail.com) Khoo EM, University of Malaya (khooem@um.edu.my) Ng CJ, University of Malaya (ngcj@um.edu.my) Conflict of Interest:None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Kumara Mendis, Senior Lecturer School of Rural Health, University of Sydney, Australia
Send e-letter to journal:
|
We admit that there is confusion regarding Table 3 as Chen has commented. The ‘Country of Publication – CY’ tag is not used anymore. The ‘Place of Publication – PT’ has the information that we obtained for the Table 3. Table 3 should read as: ‘Frequency of family medicine citations in PubMed by country of publication from 1960-2003’. It must be clear that this is not the country of author’s origin as it is implied in our article. The reason why we did not use the ‘Affiliation – AD’ tag to obtain the ‘country of the first author’ is exactly the same – it was included in only a half of the citations. Extraction of the country from the AD field is much easier as we had written the data into a RDBMS. However for more than 50% of the records it was empty (as correctly pointed out by Chen). The problem about Journal names would have been clear from our explanation given to comments by Håkansson & Kragstrup. The listing of authors is a big problem especially when authors are not consistent with especially the initial/s. This topic is dealt in great length in our next paper regarding the Australian General Practice. This has been submitted to the Medical Journal of Australia. We fully endorse Chen’s view that the ‘weak points did not’ and should not diminish PubMed’s value in the biomedical research communities. With a little bit more consistency in how we use names and key/MeSH words, we could even strengthen PubMed value. Conflict of Interest:None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Kumara Mendis, Senior Lecturer School of Rural Health, University of Sydney, Australia
Send e-letter to journal:
|
I was pleased to see comments posted to our paper by Håkansson & Kragstrup. Please accept our apologies for the delay as I saw this only on the 5 th of December. 1) Our methodology was clearly presented in the paper so that anyone could repeat it and comment. The query we used was designed after much discussions with experts in family/general practice and PubMed. Therefore this was a ‘consensus query’ so as to maximize the extraction of “family/general practice” citations from PubMed. We aimed to increase the sensitivity as much as possible while keeping out the obvious nonsense citations. We have to admit that we could not check each and every citation manually. In very simple language we asked from the query to INCLUDE the citations where ‘family practice’, ‘general practice’ or ‘family medicine’ appeared in at least (a )as one of the MeSH terms (b)in the title or (c) in the abstract. We also instructed the query to EXCLUDE any phrases like “general practice dental”. We limited the search period to 1960-01-01 to 2003-12-31 We agree that the above query could have a few ‘false positives’ as we entertained even text word such as ‘general practice’, ‘family practice’ and ‘family medicine’ anywhere in the citation. What Håkansson & Kragstrup has done is similar. What we have given in the ‘Method’ is the underlying query interpretation by the ENTREZ retrieval system of PubMed for greater accuracy. When you type like Håkansson & Kragstrup did on the PubMed search box and press Enter and then press DETAIL button to see how exactly you searched you will see what we have given as the query. Therefore we did nothing extraordinary to get what we have published. We have been technical for greater accuracy and repeatability. Håkansson & Kragstrup have NOT excluded the “general practice dental”. They may also have forgotten the search period. All these effects the number of citations retrieved.
2) Did you just forget our journal, or why was it not in your table? The methods section did not include how we obtained the journals table. We admit that was our mistake. We searched for the GP/FP journals using the PubMed “Journals database” which is separate from the main PubMed database (This can be accessed from the left side menu of PubMed). We searched for all family practice journals using the query below: “ family practice OR general practice OR family medicine OR Family Physician OR Family Practitioner” There were 21 journals cited but the ‘Scandinavian Journal of Primary Health Care’ (SJPC) was not among them. After we obtained the list of the journals, we then searched each journal using the normal PubMed query as below. Journal name [Journal] and recorded the citations for the period 1960-2003. We searched all the 21 individual journals that were on the list. Our paper contained the top 15 Journals. We did not combine the individual journal search with our definition query of general practice. After I read the comments from Håkansson & Kragstrup I searched the PubMed Journals database for “primary care” and obtained eight more journal names but the SJPC was yet not on the list. Then I further broadened to include “primary health care” and ended with two journals – the SJPC and SJPC Supplement. Håkansson & Kragstrup will now understand why we did not have the SJPC on our list of journals. We never intend to omit your important journal but it just did not get into our list by the method we used. We agree that we have to improve our method and will present the improved journal table at the end of the comments. A more proactive journal search method with regards to the medical specialty can be recommended to PubMed but again the problem may be the general medical journals. We could have avoided this problem if we analyzed the Journal list from our relation database once we wrote the MEDLINE text file to the Access RDBMS. Even if we sorted the journals table alphabetically it was a huge task going though each entry to see what journals to individually search. This is why we relied as much as possible on the direct PubMed search queries that any person could replicate or verify. We publish an updated version of our journal table as a result of questions raised by Håkansson & Kragstrup. We designed a query that joined the individual journal name AND the definition query of GP/FP (our main query). The search results are given in the table below. We have also included the percentage of GP/FP articles in each journal.
Table 2 PubMed citations for the family/general practice journals: Journal name search alone compared to journal name combined with general practice definition query
I am satisfied that I sent this article to Family Practice – the journal with the second highest percentage (61.3%) of the GP/FP discipline content. There can be other interpretations as well – that at least some of the review articles may have been incorrectly indexed. For example if an article regarding the current management of cardiac failure in general practice appears in Australian family Physician it should at least have one MeSH (key) word – such as general practice. This is entirely different from an article in Lancet about current management of cardiac failure which is undoubtedly indexed under cardiology or internal medicine. 3) Was it wise to omit the search words “Primary care OR Primary health care”? The answer is yes – we had little choice because our paper was about ‘PubMed perspective of Family Medicine’. This question had been debated at length before publication and I will give our answer again. The European definition of General Practice / Family Medicine will be acceptable for most of us including those from Asian countries. “General practice/family medicine is an academic and scientific discipline, with its own educational content, research, evidence base and clinical activity, and a clinical specialty orientated to primary care.” We have a very specific discipline oriented towards primary care which is a broader group of disciplines. Table 1 PubMed searches for primary health care, family practice and general practice
* #2 is our definition query of general/family practice PubMed search for “primary health care” resulted in 46948 citations. 41544 of them did not have the words “general practice” even as a text word anywhere in the citation. Similarly, 37322 did not have “family practice”. Therefore we do not think it would have been wise to use primary health care as a search term for the more specific discipline of general practice or family medicine. It is advisable to use the most specific word or phrase possible when we search. It would be similar to use “internal medicine” to search about “cardiology”. Furthermore there are many University Departments world over name ‘Department of General Practice and Primary Care’. This means that there is a difference in the two names at least in the academia. Otherwise it would be named using the word OR a backslash (/). More than any other reason, this was the best possible way of searching for the citations in the discipline of general practice / family medicine. Even if you identified yourself as a primary health care specialist you can very well not be a family physician. Increasingly family doctors have special interests. The mere fact that I am in a Department of general practice that does not mean that all my research publications are about general practice? Our objective within constrains of PubMed rules is to improve our data retrieval. If we could influence the National Library of Medicine authorities to bring about changes so that the discipline is more easily retrieved, it would be marvelous. However I doubt much could be done. If we identify ourselves as a family medicine specialist or as a general practitioner and agree with the European definition, and if we publish a paper that is connected to do with our discipline, please add a key word – general practice or family practice. That is the only way we can solve this problem at least the short term. PubMed will likely to be the most often used public domain medical knowledgebase for a long time to come. Kumara Mendis Conflict of Interest:None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Anders Håkansson, Professor of general practice General Practice/Family Medicine, Malmö University Hospital, SE-205 02 Malmö, Sweden, Jakob Kragstrup
Send e-letter to journal:
|
Scandinavian Journal of Primary Health Care was established as an international journal in 1983, and it has been indexed for Medline ever since. It is the leading journal of general practice / family medicine in the Nordic countries, and it is owned jointly by the societies of general practice / family medicine in Denmark, Finland, Iceland, Norway, and Sweden. In Family Practice 5/2005, we read with great interest the article by Mendis and Solangaarachchi [1], especially since one of us recently did a similar study, looking at the historical development of primary care research from a Pub Med perspective [2]. However, we were surprised not to find our own journal in the table presenting the frequency of family medicine citations by 15 leading family practice journals in Pub Med from 1960 to 2003. Since the authors did a special Pub Med search of their own, we did not try to repeat it. Instead, we did an “ordinary” Pub Med search on 5 November 2005, during our last editorial meeting. We searched the Pub Med database for “(General practice OR Family medicine OR Family practice) AND Scand J Prim Health Care” and found 440 articles. If we included “Primary care OR Primary health care”, the number of articles increased to 692. The search word “Scand J Prim Health Care” gave 1087 articles; thus only 40% and 64%, respectively, of the the total number of articles were identified using the two search strategies above. We did our first-mentioned search “(General practice OR Family medicine OR Family practice) AND Name of Journal” also for the other two leading journals of general practice / family medicine in Europe. For British Journal of General Practice (12 issues per year) we found 2353 articles, and for Family Practice (6 issues per year) we found 1151 articles; thus only about two thirds of the number of articles found by the Sri Lankan authors. This should to be compared with the above-mentioned 440 articles for Scandinavian Journal of Primary Health Care (4 issues per year); a figure that probably would place us as number 10 in the “Top-15-table” on page 572 in Family Practice 5/2005 [1]. Our questions to Mendis and Solangaarachchi are: Did you just forget our journal, or why was it not in your table? Was it wise to omit the search words “Primary care OR Primary health care”? Anders Håkansson, Professor of general practice, National editor, Scandinavian Journal of Primary Health Care, Department of Clinical Sciences, Malmö, General Practice / Family Medicine, Lund University, Malmö University Hospital, SE-205 02 Malmö, Sweden, Email: anders.hakansson@med.lu.se Jakob Kragstrup, Professor of general practice, Chief editor, Scandinavian Journal of Primary Health Care References 1. Mendis K, Solangaarachchi I. PubMed perspective of family medicine research: where does it stand? Fam Pract 2005; 22: 570–575. 2. Ovhed I, van Royen P, Håkansson A. What is the future of primary care research? Probably fairly bright, if we may believe the historical development. Scand J Prim Health Care 2005; 23: 248–253. Conflict of Interest:None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Tzeng-Ji Chen, Associate Professor Dept. of Family Medicine, National Yang-Ming University School of Medicine, Taipei 11217, Taiwan
Send e-letter to journal:
|
Thanks to Mendis and Solangaarachchi’s accurate description of their methods,[1] I could re-run the PubMed queries directly. Because the time frame was set between 1960 and 2003, the numbers of downloaded citations did not differ greatly (50337 on 13 September 2005 vs 50288 in November 2004). A detailed analysis had confirmed my suspicion. Mendis and Solangaarachchi seemed to have confused the country of origin with the place of publication. The place of publication of an article indicated the publishing journal’s country, while an author’s or her/his affiliation’s country could only be derived from the affiliation field. But, the National Library of Medicine (USA) began to catalogue the institutional affiliation and address of the ‘first’ author of an article as late as in 1987. Only 42.5% (21401) of the downloaded citations contained such a field. Moreover, the data in the affiliation field of PubMed were not indexed syntactically as those in the Thomson-ISI databases. Extracting the country names from the affiliation field of PubMed demanded extra manual efforts beside automatic processing. Totally 2619 journals appeared in the downloaded citations. Among the family practice journals, Mendis and Solangaarachchi had missed mentioning one important journal that was continued by the British Journal of General Practice in 1990: the Journal of the Royal College of General Practitioners (1614 citations) and its Occasional Paper series (45 citations). In listing the most prolific authors, one should pay attention to the namesakes in which people with different full names shared the same surname and forename initial(s). To help disambiguate author names, the National Library of Medicine started to enter the authors’ full first and middle names, if available, into the forename field of MEDLINE in 2002.[2] In the downloaded citations, Jones R represented Bob Jones, Roger Jones, and Russell Jones. The new indexing measure in PubMed was still unable to solve the problems in which different people had the same full names or a person had several name labels. As to the publication types of the PubMed citations, I would like to point out that ‘randomized controlled trials’ were only a subset of ‘clinical trials’ that included all types and phases of clinical trials. An impression leading to double counting should be avoided. The PubMed with its open access and powerful searching ability has become one of the indispensable reservoirs for bibliometric studies. Its weak points did not diminish its value in the biomedical research communities. References: 1 Mendis K, Solangaarachchi I. PubMed perspective of family medicine research: where does it stand? Fam Pract 2005; 22: 570-575. 2 Anonymous. MEDLINE® data changes – 2002. NLM Technical Bulletin 2001; 323: e11. Conflict of Interest:None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||