Article type
Year
Abstract
Introduction: Geriatric patients need specific treatments provided by doctors who choose these treatments based on existing evidence. However, evidence on many geriatric topics is hard to find. We aimed to create user-friendly search strategies with excellent operating characteristics to find geriatric evidence in Medline quickly and easily.
Methods: We used a diagnostic test framework to create the search strategies. We compared the search strategies' retrieved articles with a reference set as index test. This reference set consisted of 2255 articles from 23 biomedical journals in Medline. Each of these articles was labeled as relevant, not relevant or possibly relevant for geriatric medicine. From the relevant articles, we identified search terms to compose different search strategies. We compared the retrieved result of the different search strategies with the reference set to calculate the sensitivity, specificity, precision, accuracy and number-needed-to read (NNR).
Results: The most sensitive search strategy had a sensitivity of 92.0%, a specificity of 86.9%, a precision of 71.1%, an accuracy of 88.2% and a NNR of 1.40. The most specific search strategy had a specificity of 96.0%, a sensitivity of 69.6%, a precision of 86.1%, an accuracy of 89.2% and a NNR of 1.16.
Conclusion: Our geriatric search strategies simplify searching for relevant literature and therefore contribute to a better evidence-based practice. The search strategies are useful to both the clinician who wishes a quick answer to a clinical question and to the researcher who wants to find as many articles as possible without missing too much relevant information.
Methods: We used a diagnostic test framework to create the search strategies. We compared the search strategies' retrieved articles with a reference set as index test. This reference set consisted of 2255 articles from 23 biomedical journals in Medline. Each of these articles was labeled as relevant, not relevant or possibly relevant for geriatric medicine. From the relevant articles, we identified search terms to compose different search strategies. We compared the retrieved result of the different search strategies with the reference set to calculate the sensitivity, specificity, precision, accuracy and number-needed-to read (NNR).
Results: The most sensitive search strategy had a sensitivity of 92.0%, a specificity of 86.9%, a precision of 71.1%, an accuracy of 88.2% and a NNR of 1.40. The most specific search strategy had a specificity of 96.0%, a sensitivity of 69.6%, a precision of 86.1%, an accuracy of 89.2% and a NNR of 1.16.
Conclusion: Our geriatric search strategies simplify searching for relevant literature and therefore contribute to a better evidence-based practice. The search strategies are useful to both the clinician who wishes a quick answer to a clinical question and to the researcher who wants to find as many articles as possible without missing too much relevant information.