Skip Navigation
Department of Health and Human Services www.hhs.gov
Agency for Healthcare Research Quality www.ahrq.gov
AHRQ Home | Questions? | Contact Us | Site Map | What's New | Browse | Información en español | E-mail UpdatesE-mail Updates
 
Slide Tray
0 slides

Return to Slide Library

Slides

Add Presentation to Slide Tray Presentation:

Searching for Relevant Studies

Slides: 1–12 of 54
Searching for Relevant Studies. Prepared for: The Agency for Healthcare Research and Quality (AHRQ), Training Modules for Systematic Reviews Methods Guide, www.ahrq.gov
Add to tray Remove from tray

Searching for Relevant Studies

This figure presents an overview of the steps in the systematic review process. The first step, the preparation of the topic, requires the refinement of the topic and development of an analytic framework. The second step is to search for and select studies for inclusion, which involves identifying study eligibility criteria, searching for relevant studies, and selecting evidence for inclusion. The third step is to abstract data, which involves extracting data from individual studies. The fourth step is to analyze and synthesize studies, which involves assessing the quality of individual studies, assessing applicability, presenting findings in tables, synthesizing quantitative data, and grading the strength of evidence. The final step is to report the systematic review. This module focuses on searching for relevant studies.
Add to tray Remove from tray

Systematic Review Process Overview

Keywords: systematic review process
Learning Objectives. To substantiate why prudent literature searching is important. To describe important tenets of literature searching. To describe why transparency in the search process is important.
Add to tray Remove from tray

Learning Objectives

Keywords: literature search | transparency
On this slide, there is a long yellow line that is balanced at its midpoint on a triangle of the same color, just like a seesaw. The left side of the line indicates “Precision: the proportion of retrieved articles that are relevant.” This is sometimes also referred to as specificity. The right side of the line indicates “Recall: the proportion of potentially relevant articles retrieved by search.” This is sometimes also referred to as sensitivity.
Add to tray Remove from tray

Balancing Precision and Recall

Keywords: systematic review
Presenting Popular Studies Is Not Enough. One evaluation compared the results of popular studies against others with populations and methodologies that were similar or superior. Evaluation of 45 studies published from 1993 to 2000 in journals with high impact factors that reported positive findings and were cited more than 1,000 times. Results were compared against other studies with the same comparators by employing populations and methodologies that were similar or superior. Opposite/null findings or much more dramatic effects were found 31 percent of the time. Eighty-three percent of nonrandomized studies and 23 percent of randomized controlled trials were subsequently contradicted.
Add to tray Remove from tray

Presenting Popular Studies Is Not Enough

Keywords: systematic review | comparator | population | methodology
Where To Begin. When a general topic is proposed, it is tempting to begin by extensively searching for primary literature. Before performing the extensive literature search, it is crucial to understand the topic, devise an analytic framework, ask clearly defined key questions, and understand the scope of the review to be conducted.
Add to tray Remove from tray

Where To Begin

Keywords: systematic review | observational studies | meta-analysis | analytic framework | key question
Specialized Electronic Databases. Although the topic area dictates the databases to be included, the following are common ones: MEDLINE (General). Cochrane Central (General). On this slide, MEDLINE and Cochrane Central are connected with a bracket followed by: Generally a minimum requirement for a thorough search of literature. EMBASE (General, more international in scope). PsychLit/PsychINFO (Psychology). AIDSLine (HIV/AIDS). CINAHL (Nursing). TOXNET (Adverse Events/Toxicology).
Add to tray Remove from tray

Specialized Electronic Databases

Keywords: electronic databases
MEDLINE Alone Is Not Enough. Study 1 (Sassi et al., 2002): sensitivity and specificity of using MEDLINE versus nine databases. This study focused on ability to retrieve economic analyses from January to March 1997. MEDLINE-only searches had a sensitivity of 72 percent and a specificity of 75 percent versus nine databases. Study 2 (Betrán et al., 2005): systematic review of the prevalence of maternal mortality and morbidity from 1997 to 2002. Multiple databases were searched (MEDLINE, EMBASE, BIOSIS, LILACS). 60 percent of citations were found in more than one database. The MEDLINE search had 20 percent of nonreplicated citations, EMBASE had 7.4 percent, and LILACS had 5.6 percent.
Add to tray Remove from tray

MEDLINE Alone Is Not Enough

Keywords: electronic databases
Identifying Key Articles To Refine a Pilot Search. Perform a pilot search and compare to see if output matches articles already identified by the research group, key informants, and manual searches of references of these articles. Pilot searching is usually limited to MEDLINE and then adapted to other databases. After pilot searching, refine the search. Searching the literature is an iterative process.
Add to tray Remove from tray

Identifying Key Articles To Refine a Pilot Search

Keywords: systematic review | search strategy
Multiple Searches Help With an Extensive Literature Base (I). Multiple searches within the same database may be necessary. Efficacy and harms searches may be separate. If the available literature is small, a single broad search is appropriate. If not so, separate searches are prudent.
Add to tray Remove from tray

Multiple Searches Help With an Extensive Literature Base

Keywords: electronic database | search strategy
Multiple Searches Help With an Extensive Literature Base (II). An efficacy search may include the disease, intervention, and terms for randomized controlled trials (RCTs). Efficacy evaluations frequently are limited to RCTs. Fewer inherent biases and higher quality evidence. A harms search may be broader and include just the intervention and harms and/or disease. The paucity of harms data in published RCTs frequently requires that non-RCTs be included. Authors would likely be interested in harms regardless of the reason for the intervention.
Add to tray Remove from tray

Multiple Searches Help With an Extensive Literature Base (II)

Keywords: comparative effectiveness review | randomized controlled trial | observational studies | intervention
Boolean Operators Help To Balance Precision and Recall. “OR”—makes the search broader. “AND”—makes the search more selective. “/”—means all subheadings are searched. When a search term is entered, you can selectively choose some subheadings (such as pharmacology for pharmacology studies). Subheadings are generally imprecise; accepting all subheadings increases the yield. “adj”—cuts down on miscellaneous citations. “Heart adj failure” would find only instances in which “heart” is next to “failure.” “$”—truncates a word with different endings. “Analy$” would find words like analysis, analyses, analyze, and analyse. “.ti” or “.ab”—searches in the title or abstract for the word of interest. Can find citations where the word you search for is not a keyword. “.mp” or multiple posting—a text word search; does not need to be a key word.
Add to tray Remove from tray

Boolean Operators Help To Balance Precision and Recall

Keywords: Boolean operators | search strategy
Pages: [1] 2 3 4 5 Next
 
 
 
 
www.ahrq.gov AHRQ Home | Questions? | Contact AHRQ | Site Map | Accessibility | Privacy Policy | Freedom of Information Act | Disclaimers
U.S. Department of Health & Human Services | The White House | USA.gov: The U.S. Government's Official Web Portal

Agency for Healthcare Research and Quality • 540 Gaither Road Rockville, MD 20850 • Telephone: (301) 427-1364