| systematic | | ______ effects means every study is estimating different tx effects - addresses issue of heterogeneity |
| sensitivity | | if p values are ____ than 0.05, the less likely that the differences observed btwn studies are due to chance alone and that heterogeneity btwn studies exist |
| hypothesis | | if the CI includes ___ results aren't significant for OR or RR |
| meatanalysis | | the research ___________ is the question that's being answered, which should be a FOCUSED question; should be clear to the reader that previous studies are inconclusive or otherwise limited |
| search | | the ______-_______ test combines the OR or RR of some event occurring across studies using a weighed average w/in study results |
| fixed | | study ________ is evaluated and rated by investigators; poor studies overestimate tx effects; may determine cutoff score for selecting studies for pooling |
| evaluating | | _____ effects are when every study is estimating the same tx effect (OR) - assumes homogeneity |
| quantitative | | this determines if data from various studies should be meaningfully combined by looking at study pop, diagnostic criteria, tx regimens and endpoints; means they aren't similar enough and pooling data may be problematic |
| methods | | a ____-________ is performed to increase STATISTICAL POWER for primary endpoints by pooling data to INCREASE sample size, to resolve UNCERTAINTY that may exist about a topic, to improve estimates of EFFECT SIZE - magnitude of difference found in sample which more clearly defines the "true" difference btwn sample and control, and to answer question not posed at start of individual trials |
| quality | | this type of review article is tertiary literature, provides a general overview on a topic, may not have all studies on a topic included |
| extraction | | varias ______ strategies include looking at the computer, unpublished data, personal communication; don't only use computer to avoid publication bias; authors should tell you how it was conducted |
| greater | | ____________ systematic reviews, or meta-analysis, is a comprehensive "study of studies;" combines and statistically POOLS data from previous research; can be considered primary literature since it may provide "new knowledge" on a topic |
| publication | | ______ plots or petro diagrams shows point estimates and 95% CI in graphical form |
| heterogenous | | if the p value is _______ than 0.05 it means the studies are homogeneous and can be pooled |
| forest | | ___________ analysis determines how sensitive the results are to changes in the way meta-analysis was conducted |
| mantel-haenzel | | ___________ systematic reviews have a DEFINED METHODOLOGY but data from studies are NOT POOLED; there is a specific question that is addressed; there's an explaination of how search was conducted; there's a describtion of what studies were included and excluded |
| inclusionexclusion | | the inverted funnel plot tests for ___________ bias; identifies outlier studies; compares studies that favor outcome and control |
| one | | studies should be selected based on _______ and not results to avoid selection bias |
| homogenous | | _________/_________ criteria list the criteria for studies to be included in analysis; list of trials included should be available for reader to review; should be established before meta-analysis conducted |
| random | | this determines if data from various studies should be meaningfully combined by looking at study pop, diagnostic criteria, tx regimens and endpoints; means they can be pooled if similar enough |
| qualitative | | this type of review article is focused on a specific topic, has a predefined research methodology, rigorous critical appraisal of studies is included, may be either QUALITATIVE (summarizes results) or QUANTITATIVE (meta-analysis) |
| narrative | | more than one person should be involved in data __________ to avoid observation/interpretation bias; at least 2 people do this to data to make sure nothing is missed or not interpreted correctly |
| less | | when ___________ reviews, one must look at why it was written, how information was obtained, what the criteria for study inclusion was, evidence vs. opinion (ALWAYS check references), if the conclusion is consistent w/the information, and how can the information be applied |
