Nutrition Research Ch. 12
Each review follows a rigorous process to ensure ____ and minimize ____ . In the process of creating the EAL, the Academy established guidelines for conducting systematic reviews, which generally take about ___ to ___ months to complete.
-transparency -bias -12 to 18
larger studies should carry more ____. The method typically used in meta-analysis to weight studies is called ____ ___ ____, which uses the standard error to calculate a weight for each study.
-weight - inverse-variance weighting
The systematic review process for the EAL includes the following five steps.
1) Determine the evidence analysis question 2) Gather and classify the evidence (data). 3) Critique each article for risk of bias. 4) Summarize the evidence. 5) Write the conclusion statement and assign grading. (p.333)
In a fixed effects model, it is assumed that the studies have been conducted under ____ conditions with ____ participants. When studies are more varied, researchers often use a ____ ____ model, which adjusts for the variability between the studies.
-similar -similar -random effects
Reading the results of a meta-analysis can be fun because the ___ and the ___ ___ are often displayed using a __ ___
-studies -summary estimate -forest plot
Systematic reviews are not primary research; they belong in the category of ____ research because they collect and distill information from both primary and secondary sources. Secondary research includes _____ reviews in which authors evaluate, interpret, and summarize evidence from primary studies in a particular research area.
-tertiary -narrative
Systematic reviews normally include ____ studies with similar research designs, such as RCTs. If the research studies are qualitative, researchers can do a ____ ____.
-quantitative -meta-synthesis
Meta-analysis is a statistical technique used to combine the ____ of a number of studies, so you can see the size of an ____ of an intervention or treatment from multiple studies on a ____ outcome
-results -effect -clinical
To compare and summarize the results in a meta-analysis when the outcome is dichotomous, researchers often use ____ or ____. Both measure ___ ___
-risk ratios or odds ratios -effect size
Systematic review is a ... It may or may not include a quantitative synthesis called a ____ ____, which is a statistical technique that combines the results from separate studies
-scientific investigation that focuses on a specific question and uses explicit, planned scientific methods to identify, select, assess, and summarize the findings of similar but separate studies. -meta-analysis
Narrative reviews are different from systematic reviews in a number of ways:
-The research question in a narrative review tends to be broad in scope; whereas in a systematic review the question is more focused. -Systematic reviews involve a more comprehensive literature search than narrative reviews and often including unpublished research to reduce publication bias. -Studies used in a systematic review are more rigorously appraised than are most narrative studies. -Narrative reviews include a qualitative summary, whereas systematic reviews usually include a quantitative summary of the literature. -All decisions in a systematic review, from picking to appraising studies, are more explicit and thorough than in most narrative studies.
mean difference (MD) -used when -shows -estimates
-This is used when all studies in the meta-analysis measure the outcome on the same scale. -The mean difference shows the absolute difference between two groups. -It estimates how much the intervention changed the outcome on average compared with the control group.
standardized mean difference (SMD) -used when -does what -reported as -calculated by -end up with -as it increases...
-This is used when different scales are used to measure the same outcome in the studies. -standardizes the results of the studies to a uniform scale so you can compare studies. -often reported as Cohen's d or simply d -generally calculated by taking the difference in mean outcome between the two groups (experimental minus control) and dividing it by the pooled standard deviation (a correction factor may be used). -What you end up with is a number that shows how far apart the means are in standard deviation units. -As the SMD increases, the difference between the intervention and control groups increases.
Bias is ... Bias is not a random error that shows up haphazardly; it is a ____ error.
-anything that influences a study in such a way that it distorts the findings -systematic
I2 describes the percentage of variation across studies that is not due to ____. An I2 of ___ to ___% may represent substantial heterogeneity. If results for the studies are very heterogeneous, a meta-analysis is ____ likely to be appropriate. When heterogeneity is low, a ___ ___ model can be used.
-chance -50 to 90 -not - fixed effects
A meta-synthesis involves ... In meta-synthesis, the researchers rigorously examine and interpret the findings of research studies using a ____ method.
-compiling, interpreting, and integrating qualitative studies to possibly build a theory, explain a theory, or for some other purpose. -qualitative
Heterogeneity in a meta-analysis refers to the variation in the ____ and ____ of the effect size for the studies. Ideally, the results of the different studies should be _____. Researchers often do a statistical test, such as ___ , to estimate the level of heterogeneity.
-direction and magnitude -similar -I2
In meta-analysis, researchers combine the different effect size estimates from each study to give an overall best estimate of the size of the ____. The effect size is often one of the following: mean difference or standardized mean difference for ____ data, and risk ratio or odds ratio for ____ data.
-effect -continuous -binary
Effect size is the ____ of an effect. It could be the ____ of the difference in an outcome measure between the experimental and control groups in an experimental study. It could also be the ____ of the relationship between two variables, as seen, for example, in correlation
-magnitude -size -strength
Effect sizes are used to compare and summarize the results of different studies in a meta-analysis. We use ____ ___ and ____ ___ ___ when comparing studies with continuous outcomes (meaning ___ or ____ data)
-mean difference -standardized mean difference -interval or ratio
When you want to examine the size of the difference in means between two groups in a study, all you have to do is subtract the ____. This difference is called the ___ ___ and is an example of an ____size.
-means -mean difference -effect
If a study's results show an RR of 1.0, that means there is ___ difference in risk between the groups and the exposure did not increase or decrease risks of the outcome. If the RR is ___ than 1, it suggests an increased risk of that outcome in the exposed group. If the RR is ____ than 1, it suggests a reduced risk in the exposed group.
-no -greater -less
Whereas relative risk tells you about probability, the odds ratio (OR) tells you about the ____ of an outcome. Odds ratios are ___ ___ or ____ an outcome. An OR of 1.0 means that the odds of an outcome in each group are the ____. If the OR is ____ than 1, then exposure is positively related to an outcome and could be a causal factor. If the OR is ____ than 1, exposure is negatively related to disease, meaning the exposure is protective.
-odds -associated with or predict -same -greater -less
Systematic reviews critically appraise and synthesize the results of ____ research to answer health care-related questions. Systematic reviews inform and serve as the foundation for ...
-primary -evidence-based practice guidelines
10 steps for a lit review
1) Formulate the research questions and the purpose of the review. 2) Define inclusion and exclusion criteria for studies you will use 3) Develop a search strategy, including selecting relevant electronic databases and keywords to be used, as well as other strategies 4) Conduct the search of the literature using your search strategy. 5) Screen the records. 6) Assess the full text of the selected studies from the previous step for eligibility using reviewers. 7) Critically assess the risk of bias in the included studies to help determine the strength of the evidence based on the quality of the studies 8) Put together the studies' findings to yield a result; this is called evidence synthesis 9) Summarize and discuss the findings from the synthesis. 10) Finish the final report for publication
6 domains of Cochrane Collaboration tool
1. Random sequence generation (selection bias) 2. Allocation concealment (selection bias) 3. Blinding of participants and personnel (performance bias) 4. Blinding of outcome assessment (detection bias) 5. Incomplete outcome data (attrition bias) 6. Selection outcome reporting (reporting bias)
6. Selection outcome reporting (reporting bias)
A study's original protocol spells out the outcomes that will be reported and how they will be reported. Bias may occur if, for instance, an outcome from the protocol is not reported at all or is reported incompletely, or if the outcome is measured differently than originally stated.
4. Blinding of outcome assessment (detection bias)
If the people responsible for assessing patient outcomes know which intervention a participant received, they could bias how the outcome was measured. Blinding of outcome assessment means that the people assessing patients do not know which intervention patients received.
2. Allocation concealment (selection bias)
Random allocation concealment involves implementing the random sequence in a way that conceals group placement from anyone who enters participants in a study. If people who enter participants know, or can detect, the upcoming allocations (such as the next participant will go into the experimental group), they may channel participants to certain groups.
clinical practice guidelines are...
Statements that include recommendations intended to optimize patient care that are informed by a systematic review of evidence and an assessment of the benefits and harms of alternative care options.
The expert work group who worked on the systematic review also develops the nutrition practice guidelines using the following process
Step 1. Clinical Algorithm: The expert work group develops a clinical algorithm for the guideline based on the Academy's Nutrition Care Process. In this case, algorithms are illustrated in step-by-step flowcharts to illustrate how each recommendation can be used to treat patients. Step 2. Draft Recommendations: The expert work group prepares draft recommendations for the guideline from evidence analysis using an expert consensus method. The first statement in a recommendation explains what the dietitian should do (or not do), and the next statement explains why. More than one recommendation is often needed. Each recommendation is rated based, as usual, on the strength of the evidence. Each rating is also given a label: conditional or imperative. Imperative recommendations have no conditions attached to their use, but conditional recommendations use an "if, then" format so you know in which situations it can be used. Step 3. Complete Writing Guideline: Each guideline for a specific disease has a format: Introduction, Major Recommendations, Background Information, and References/Appendices. Step 4. Internal and External Review: The guidelines are reviewed internally and externally using a standardized evaluation tool (Appraisal of Guidelines for Research and Evaluation). EAL guidelines are reviewed for each topic every 5 years. Step 5. Approval and Publish: The Evidence-Based Practice Committee must approve each guideline, and then it is published on the EAL
1. Random sequence generation (selection bias)
The method of assigning participants to groups should be fully described to determine if it produced comparable groups. To randomly put participants in groups first involves generating an unpredictable random sequence. Keep in mind that for many clinical trials participants are entered into the trial over a period of time. As patients come in and are judged eligible for the study, they are assigned to a group using the random sequence.
5. Incomplete outcome data (attrition bias)
This domain looks at how attrition and exclusions were handled by the researchers and whether missing data (due to dropouts, etc.) could have affected the results.
3. Blinding of participants and personnel (performance bias)
This looks at the measures used to blind participants and study personnel from knowing which intervention a participant received.
A necessary step in a systematic review is to examine each of the included studies for ____.
bias
Risk ration/relative risk
compares the probability of an outcome (usually a disease) in two groups. -One group is exposed to something, such as a risk factor, and the other group is not exposed to it. -Relative risk expresses how much more (or less) likely it is for an exposed person to develop an outcome (relative to an unexposed person). -Exposure is often a risk factor for a disease, but it could also be a treatment.
relative risk works well as an effect size because it gives both ___ and ____ of the effect.
direction and magnitude
know how to read one
p 330-332
Mixed methods systematic review is a synthesis of findings from both ____ and ____ studies that brings together research knowledge.
quantitative and qualitative
A number of researchers use a Cochrane Collaboration tool to assess risk of bias that includes these ____ domains as well as a miscellaneous domain (other sources of bias)
six
Work group members are subject matter experts in their assigned topic and must meet the following criteria to be part of an EAL work group team:
• Minimum of 5 years experience in practice and/or research. • At least 3 years of work experience related to the topic. • Advanced degree in the topic area or 8 years of experience. • Willingness to volunteer their time with no compensation for their contribution to the work group project. • Willingness to disclose potential conflicts of interest. • Attend an Academy orientation on the systematic review process.