Reader questions about the pancreatitis case
We received a few questions from a group of readers, sent to us by information specialist Ani Orchanian-Cheff, about the April JMLA case on diagnosing acute pancreatitis. Their questions and responses from the case's librarian co-authors, Julie Beauregard and Jennifer Lyon:
We were wondering how many citations you went through in order to find the 4 you ended up summarizing. Some of us thought that you might have gone through the ninety results from your text word search strategy on page 123, while others thought perhaps you went through a combination of the results of all three searches on that page.
We looked at a combination of the results of multiple searches, weeding through the abstracts in the online literature databases to select a smaller number of articles to look at in full-text. We pulled the full-text of about 40 articles. Those were skimmed and pared down based on article type, number of patients, presence of a gold standard for diagnostic criteria and sound methodology -- this process resulted in a smaller number, approximately 10 items, which we read more closely, weighing both their methodology and their relevance to the case at hand. Based on those factors, we chose the final 4 articles to summarize.
A related question was which of your searches was your final/definitive search, or was your final search a combination of the various MeSH and textword suggestions from the search possibilities you outlined. Would you mind sharing it with us?
We used multiple searches with various MeSH terms and textwords and combined the results of those searches, using citation management software (Reference Manager) to eliminate duplicates and organize the results. Between us, we tried over half-a-dozen search combinations and the strategies presented in the article are a synthesis of the most useful elements of the various strategies we tested.
It was interesting for us to see differences in practice patterns. For example, some of us are more apt to point clinicians toward existing guidelines and evidence summaries (such as the Best Evidence Topic Report on the same topic in EMJ), rather than recommending search and appraisal from scratch as you have done.
Our standard practice is to use the primary literature directly as much as possible, unless we receive a direct request for practice guidelines. Often, we will include a guideline or representative review article along with our own summary of the primary data if we feel it is reliable and helpful. However, we sometimes find that practice guidelines or other evidence-summary materials may not fully represent the related primary data or may omit some details of the original research that are relevant to the case at hand. We also find that a thorough consideration of the primary literature proves very useful in gauging the quality and comprehensiveness of review articles and guidelines.
Approximately how long does it take you to do this appraisal and present it in its finished form to the residents?
The time it takes to complete the literature searching, article selection, and summarization process varies from question to question. Some questions have only limited literature available (in which case the search usually is the most time-intensive portion of the process), while others require looking at a larger number of articles (making the article selection and summarization process more time-consuming). Our average time per question ranges from 2.5-3 hours to 12+ hours depending on the complexity and volume of the literature.
We welcome reader questions and comments on the cases via the comments feature on the blog or by email (jmlaeditor@vanderbilt.edu).
We were wondering how many citations you went through in order to find the 4 you ended up summarizing. Some of us thought that you might have gone through the ninety results from your text word search strategy on page 123, while others thought perhaps you went through a combination of the results of all three searches on that page.
We looked at a combination of the results of multiple searches, weeding through the abstracts in the online literature databases to select a smaller number of articles to look at in full-text. We pulled the full-text of about 40 articles. Those were skimmed and pared down based on article type, number of patients, presence of a gold standard for diagnostic criteria and sound methodology -- this process resulted in a smaller number, approximately 10 items, which we read more closely, weighing both their methodology and their relevance to the case at hand. Based on those factors, we chose the final 4 articles to summarize.
A related question was which of your searches was your final/definitive search, or was your final search a combination of the various MeSH and textword suggestions from the search possibilities you outlined. Would you mind sharing it with us?
We used multiple searches with various MeSH terms and textwords and combined the results of those searches, using citation management software (Reference Manager) to eliminate duplicates and organize the results. Between us, we tried over half-a-dozen search combinations and the strategies presented in the article are a synthesis of the most useful elements of the various strategies we tested.
It was interesting for us to see differences in practice patterns. For example, some of us are more apt to point clinicians toward existing guidelines and evidence summaries (such as the Best Evidence Topic Report on the same topic in EMJ), rather than recommending search and appraisal from scratch as you have done.
Our standard practice is to use the primary literature directly as much as possible, unless we receive a direct request for practice guidelines. Often, we will include a guideline or representative review article along with our own summary of the primary data if we feel it is reliable and helpful. However, we sometimes find that practice guidelines or other evidence-summary materials may not fully represent the related primary data or may omit some details of the original research that are relevant to the case at hand. We also find that a thorough consideration of the primary literature proves very useful in gauging the quality and comprehensiveness of review articles and guidelines.
Approximately how long does it take you to do this appraisal and present it in its finished form to the residents?
The time it takes to complete the literature searching, article selection, and summarization process varies from question to question. Some questions have only limited literature available (in which case the search usually is the most time-intensive portion of the process), while others require looking at a larger number of articles (making the article selection and summarization process more time-consuming). Our average time per question ranges from 2.5-3 hours to 12+ hours depending on the complexity and volume of the literature.
We welcome reader questions and comments on the cases via the comments feature on the blog or by email (jmlaeditor@vanderbilt.edu).
posted by Becky at 9:59 AM
0 Comments:
Post a Comment
<< Home