Final Flashcards
What kind of review articles are narrative and comprehensive reviews
qualitative and nonsystemic
Purpose of review articles
present truth found with conflicting and variable primary literature
discuss treatment of disease states or clinical aspects of drug therapy
What kind of review articles are systematic review
qualitative and systematic
What kind of review articles are meta analysis
quantitative
What are narrative reviews
not required to conform to lit retrieval and eval process
may include authors opinion in addition to a small selection of supporting data
no reporting of methodology
typically solicited from recognized experts
What are analytic or comprehensive reviews
follow a structured approach to lit retrieval and eval that may be replicated but not entirely systematic (gray lit)
may include data beyond RCTs
should report methodology, not required
no stat analysis
What are systematic reviews
follow structured approach to lie retrieval and eval that must have the ability to be replicated
commonly includes a meta-anylsis of data
RCTs with comparable outcomes are included
What is Prisma
reporting for systematic and meta analysis
has a checklist reviewed be editorial staff
has a flow diagram published as “methodology” action of a manuscript
What are meta-anaylses
analytical technique to stat combine and weigh data from multiple studies to provide a quantitive and objective assessment
-commonly accompanies a systematic review
-may stand as an independent work
What scenarios are meta-anaylses useful for
previously conducted studies are inconclusive or contradictory
previous studies are underpowered or have small sample sizes
assessment of the occurrence of rare events
provision of guidance on a clinical problem with minimal evidence
seek to answer a different research question than the one postulated in the original trial
Important considerations for meta-analyses
must clearly define a research question with a measurable outcome
-med safety, efficacy, health related QOL
must clearly define intended study population
must define inclusion criteria prior to search
-RCTs, other primary literature
commonly rate studies based on quality of evidence
-cochrane, GRADE system
What is publication bias
the publication or non-publication of research findings depending on the nature and direction of the results
What is time-lag bias
rapid or delated publication depending on results
What is language bias
choosing to publish in a language to the benefit or detriment of a population
What is citation bias
over-or-under citation for a research work
What is multiplicative bias
publishing the same work multiple times
What is location bias
accessibility of results
What is selective bias
decision from authors to report, or not report, some outcomes and not others
What are the methods for conducting a meta-anyalsis
comparison of 2 treatments to determine the treatment of one intervention or exposure to a control population
-pair-wise analyis
compare multiple treatment to rank a network
-network analysis (requires >2 tx)
What are the issues with meta-anaylses
Publication bias
-must be addressed by authors through a funnel plot
Presence of heterogeneity
-differenced w/ included studies impact results (only study design)
Limits the ability to determine applicability of results
What is a funnel plot
scatterplot of tx vs effect sample size in adult studies
variability in tx effect should decrease as sample size increases
Statistical analysis
results typically presented with forest ploy
-odds ratio w/ 95% CI for each study
-interpret as stat sign if CI does not cross 1
-larger box=;arger number of pt experiencing event
What is the low, moderate, and high percentage that represents the amount of heterogeneity
low: <25%
moderate: 25-75%
high: >75%
represented as I2 values (I2 >40% are significant)
as primary evidence increase over time, systematic reviews and meta-analyses have come under what scrutiny
scope of clinical problem may not be understood or unable to be quantified
represent data and evidence for a set period of time
unable to easily revise based on new or conflicting data