An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
Editor—Over the past two academic years colleagues and I have used the concept of an evidence based case report, similar to that presented by Glasziou,1 in teaching evidence based medicine in the penultimate year of our undergraduate course in epidemiology and public health.
Students are asked to submit a case report based on the management of a single patient encountered during their clinical work. They are advised to identify one key intervention in the management of the case and to summarise any evidence that supports this intervention. The submission should not exceed 1000 words, of which not more than 300 should describe the clinical details of the case. Students are advised to use no more than five references and to take care to select key papers; they must describe the Medline search strategy that they used.
The case report contributes towards the students’ mark in epidemiology and public health at the end of the year. The exercise is designed to help the students relate the theory of evidence based medicine to the reality of everyday clinical practice. Marking the case reports provides good feedback on the effectiveness of our teaching in evidence based medicine. A high proportion of students display evidence of critical reading of the key references. Students tend, however, to focus on the evidence for pharmacological interventions rather than on other forms of treatment or diagnostic strategies. They are reluctant to reflect critically on the management of the case and have difficulty with formulating good questions for evidence based medicine. These observations have prompted a review of methods and content in our teaching, and we hope that the BMJ’s evidence based case reports will become a valuable teaching resource.
In this medical school, as elsewhere, we are reviewing the undergraduate curriculum. We aim to promote reflective, self critical practice combined with an understanding of the scientific method. The standard of evidence based case reports submitted by our students may emerge as a useful outcome measure for this aspect of the curriculum.
Footnotes
i.perry@ucc.ie
References
1.Glasziou P. Twenty year cough in a non-smoker. BMJ. 1998;316:1660–1661. doi: 10.1136/bmj.316.7145.1660. . (30 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 1998 Nov 14;317(7169):1386.
Results of search strategy should be given for readers
Editor—In Glasziou’s evidence based case report I found the description of the literature search too tidy for belief.1-1 My own experience as a librarian over 13 years is considerably closer to that of Naidoo—“many of the citations retrieved in a Medline search are irrelevant.”1-2 Using Medline back to 1966, I reproduced the literature search given by Glasziou and found some interesting variations on what was described in the report (table). The five year limit would obviously have excluded two of the four references that Glasziou cited as being found by the search strategy given (those published in 1989 and 1981). When I used the five year strategy I found considerably more than four references, the bulk of which Glasziou obviously considered irrelevant. If the search had gone back to 1981 (which would have retrieved the four references cited) Glasziou would have had around 100 titles to wade through in order to identify the four pertinent articles.
Table.
Number of references found with various search strategies
I realise that brevity is needed in the BMJ, but brevity can end up by being misleading. Searches are rarely as simple as that described by Glasziou, there is almost always unwanted material, and narrowing down the parameters of the search is not as easy as Glasziou suggested. As Glasziou went into so much detail regarding the search strategy, surely it would not have taken up too much extra space to for him to say “the search retrieved X references, four of which were considered relevant.” This would have introduced a note of realism into the search process—something that is extremely important when people are using unfamiliar technology and systems for the first time to look for the answer to their problems. As Naidoo has already found, answers do not always exist and are certainly not always easy to find.
References
1-1.Glasziou P. Twenty year cough in a non-smoker. BMJ. 1998;316:1660–1661. doi: 10.1136/bmj.316.7145.1660. . (30 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
1-2.Naidoo N. Information “Nuggets” are not easy to find quickly. BMJ. 1998;316:1676. doi: 10.1136/bmj.316.7145.1676. . (30 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
Editor—Glasziou has written an evidence based case report about a case in which he did a selective search on Medline to inform his decisions on treatment.2-1 We are worried that this article may convey a misleading message about evidence based medicine.2-1 All diagnostic efforts focused on cough (his patient’s problem), which subsided with empirical treatment for oesophageal reflux; nevertheless, no further evaluation for possible severe complications of gastro-oesophageal reflux is mentioned.
Adenocarcinoma of the oesophagus is among the types of cancer whose incidence in Western countries is rising fastest.2-2 Barrett’s oesophagus is found in 15% of patients with chronic gastro-oesophageal reflux and has been associated in prospective studies with up to a 125-fold excess risk of adenocarcinoma; gastro-oesophageal reflux has been associated with an increased probability of adenocarcinoma (odds ratio 2.7 (95% confidence interval 1.5 to 4.9)) after adjustment for potential confounders.2-3 Antireflux treatment controls symptoms and oesophagitis, but Barrett’s oesophagus generally persists. Evidence indicates that endoscopy should be performed in patients with long term symptoms (>5 years)2-4; on the other hand, endoscopy significantly influences doctors’ decisions about medical management of gastro-oesophageal reflux disease.2-5 We obtained this information from a search on Medline, 1996-8, using “gastro-oesophageal reflux,” “management,” and “endoscopy” as research terms (which yielded 29 articles and four close matches).
Evidence based medicine should not be regarded as a form of magic whereby solutions become evident after patients’ complaints are entered in detail on a computer; rather, it is a valuable tool that has to be used wisely by skilled clinicians.
References
2-1.Glasziou P. Twenty year cough in a non-smoker. BMJ. 1998;316:1660–1661. doi: 10.1136/bmj.316.7145.1660. . (30 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
2-2.Hansen S, Wiig JN, Giercksky KE, Tretli S. Esophageal and gastric carcinoma in Norway 1958-1992: incidence the trend variability according to morphological subtypes and organ subsites. Int J Cancer. 1997;71:340–344. doi: 10.1002/(sici)1097-0215(19970502)71:3<340::aid-ijc5>3.0.co;2-y. [DOI] [PubMed] [Google Scholar]
2-3.Chow W, Finkle WD, McLaughlin JK, Frankl H, Ziel HK, Fraumeni JF. The relation of gastroesophageal reflux disease and its treatment to adenocarcinomas of the esophagus and gastric cardia. JAMA. 1995;274:474–477. [PubMed] [Google Scholar]
2-5.Ellis KK, Oelke M, Helfand M, Lieberman D. Management of symptoms of gastroesophageal reflux disease: does endoscopy influence medical management? Am J Gastroenterol. 1997;92:1472–1474. [PubMed] [Google Scholar]
Editor—Searching for “nuggets” is not easy: it requires both skills in clinical epidemiology and in literature searching and clinical skill to assess the relevance of what is found. The exercise that Perry describes for medical students is no longer a peripheral luxury but will be central to the information-overloaded but evidence-starved practitioner of the future. Fortunately, others are easing the work of tracking down the evidence. For example, if I did the search for evidence on chronic cough today I would try the “clinical queries” section of the National Library of Medicine’s PubMed (www.ncbi.nlm.nih.gov/PubMed/clinical.html). Typing in “chronic cough” and choosing “diagnosis” (which pre-filters diagnostic articles) gives 21 “hits,” including three of the four articles I previously identified. Finding articles on differential diagnosis and on diagnosis, however, is difficult compared with finding controlled trials of treatment—now gathered in the Cochrane Controlled Trials Registry.
When I must use Medline I generally start with the last five years; only later do I expand the search to all years—as in the search about chronic cough. My results were similar to those of Bates, but many articles could be discarded after scrutiny of the titles and most after scrutiny of the abstracts. If space is limited, how much detail should be provided on the methods and results of the literature search compared with on the patient and the question, the assessment of the validity of the identified articles, and their applicability to the patient?
Finally, I agree with Zuccalcalà et al that the practice of evidence based medicine still requires clinical knowledge and wisdom in addition to information skills. Part of what is needed is the translation of epidemiological information into terms relevant for individual decision making. For example, the incidence of oesophageal adenocarcinoma is around 2/100 000 a year.3-1 If oesophageal reflux increases this by an odds ratio of 2.7 then the incidence becomes 5.4/100 000 a year. Does that warrant an endoscopy? Of course, Barrett’s oesophagus may also be detected—in 1% of older people and 3-5% of those with gastro-oesophageal reflux.3-2 Should they all be identified (including the 10-25% with asymptomatic reflux3-3)? There is no evidence that antireflux treatment reduces the risk of adenocarcinoma (though there are no controlled studies of this), and hence long term monitoring by repeat endoscopy is the only current intervention. When I next see my patient I will ask her opinion about this absolute risk and the management options.
References
3-1.Pera M, Cameron AJ, Trastek VF, Carpenter HA, Zinsmeister AR. Increasing incidence of adenocarcinoma of the esophagus and esophagogastric junction. Gastroenterology. 1993;104:510–513. doi: 10.1016/0016-5085(93)90420-h. [DOI] [PubMed] [Google Scholar]
3-2.Cameron AJ. Epidemiology of columnar-lined esophagus and adenocarcinoma. Gastroenterol Clin North Am. 1997;26:487–494. doi: 10.1016/s0889-8553(05)70308-3. [DOI] [PubMed] [Google Scholar]
3-3.Ter RB, Castell DO. Gastroesophageal reflux disease in patients with columnar-lined esophagus. Gastroenterol Clin North Am. 1997;26:549–563. doi: 10.1016/s0889-8553(05)70313-7. [DOI] [PubMed] [Google Scholar]
