Article type
Year
Abstract
Introduction: To date, few comparative reports of search methods exist for the field of HIV prevention. The U.S. Centers for Disease Control and Prevention (CDC) HIV Prevention Research Synthesis (PRS) project's ongoing database of prevention trials and related literature permits the systematic assessment of the relative productivity of specific search methods in identifying rigorous trials
Objectives: To review and compare effective methods for manual and automated identification and retrieval of evidence-based HIV prevention literature.
Methods: Manual search techniques, including systematic hand searching of health, medical, and social science journals and books, and standardized automated database searches in MEDLINE, AIDSLINE, PsycLIT, Sociofile, ERIC, and CINAHL were conducted. Searches covered the period 1988 to 1997. All retrieved citations were compiled in the PRS database. Citations with outcome findings from an intervention study were assessed for PRS relevance and methodological rigor by two independent research scientists. Inter-rater reliability was determined using percentage agreement. Sensitivity of search methods was compared, as was the productivity of individual automated databases in retrieving unique citations.
Results: Analysis of the search methods used to retrieve the first cohort of relevant trials yielded the following results: (1) AIDSLINE was the most productive of the computerized database sources; (2) hand searching of journals was most productive for retrieving unique outcome citations; and (3) hand searching and literature reviews were more productive methods than automated database searching for retrieving citations with the most centrality (e.g., relevance and methodological rigor) for the project.
Discussion: These findings indicate that manual search methods are valuable tools for retrieving HIV prevention trials for systematic reviews. No single method appears to function as a stand-alone source of information. Thus, when conducting a comprehensive search of literature, more than one method should be utilized. Overall, these findings will be useful for evidence-based researchers as well as practitioners and health care consumers as they seek to identify relevant HIV prevention research.
Objectives: To review and compare effective methods for manual and automated identification and retrieval of evidence-based HIV prevention literature.
Methods: Manual search techniques, including systematic hand searching of health, medical, and social science journals and books, and standardized automated database searches in MEDLINE, AIDSLINE, PsycLIT, Sociofile, ERIC, and CINAHL were conducted. Searches covered the period 1988 to 1997. All retrieved citations were compiled in the PRS database. Citations with outcome findings from an intervention study were assessed for PRS relevance and methodological rigor by two independent research scientists. Inter-rater reliability was determined using percentage agreement. Sensitivity of search methods was compared, as was the productivity of individual automated databases in retrieving unique citations.
Results: Analysis of the search methods used to retrieve the first cohort of relevant trials yielded the following results: (1) AIDSLINE was the most productive of the computerized database sources; (2) hand searching of journals was most productive for retrieving unique outcome citations; and (3) hand searching and literature reviews were more productive methods than automated database searching for retrieving citations with the most centrality (e.g., relevance and methodological rigor) for the project.
Discussion: These findings indicate that manual search methods are valuable tools for retrieving HIV prevention trials for systematic reviews. No single method appears to function as a stand-alone source of information. Thus, when conducting a comprehensive search of literature, more than one method should be utilized. Overall, these findings will be useful for evidence-based researchers as well as practitioners and health care consumers as they seek to identify relevant HIV prevention research.