Research Methods Exam 3

Epidemiology

study of the distribution of diseases, and its determinants, and the application of this study to control health problems 

Nutritional epidemiology

examines how diet and nutrition affect health outcomes in populations, exploring links between dietary patterns and diseases like obesity, cardiovascular disease, and cancer

What is an exposure?

refers to dietary components, nutrients, foods, or behaviors, that may influence disease risk (IV)  

What is an outcome?

usually refers to a disease or condition (DV) 

What does frequency in epidemiology include?

prevalence and incidence

What does prevalence measure in epidemiology?

the number of people who have the disease at a specific point in time

What does incidence measure in epidemiology?

the number of people who get the disease over a specified period of time, the number of new cases

What does association measure in epidemiology?

if there is more disease among the exposed group compared to the non-exposed group

What are common epidemiological research designs?

case-control, cohort, and cross-sectional

Case-control study

selects patients on the outcome or disease and compares the exposure in this group to a very similar group of people who don’t have the outcome or disease  

Cohort study

follows large group of people over a long period of time, years or decades, to find association of an exposure with disease outcomes  

Cross-sectional study

exposure and outcome are measured at the same time, lack of time dimension limits the ability to make causal inference

What is the weakest epidemiological study design?

cross-sectional

Prospective cohort study

goes forward in time with the cohort (outcome hasn’t occurred yet) 

Retrospective cohort study

analyzes existing data to compare outcomes between groups exposed and not exposed to a specific risk factor in the past  

Relative risk / odds ratio

measures the strength of association between a dietary factor and a health outcome

What does a RR of 1 indicate?

there is no difference in risk between the groups

What does a RR > 1 indicate?

suggests increased risk of the outcome in exposed group

What does a RR < 1 indicate?

suggests the exposure protects participants from the outcome

How is RR calculated?

(relative risk - 1) x 100

What are cases in a case-control study?

participants with the disease

What are controls in a case-control study?

participants without the disease that are identical to the cases

Odds ratio

the odds that a case has been exposed to the risk factor of interest compared to the odds that a control has been exposed to the same factor

What does an OR = 1 indicate?

odds of exposure are the same for cases and controls

What does an OR > 1 indicate?

exposure is positively associated with the disease and could be a causal factor

What does an OR < 1 indicate?

exposure is negatively associated with the disease, could be a protective factor

What does a small confidence interval indicate for RR/OR?

a higher precision of the RR or OR

What is the focus in case-control studies?

the outcome

What is the focus in retrospective cohort studies?

the researchers separate groups by the exposure

What does a correlation measure?

the strength and direction of a linear relationship between two variables

What does a correlation of +1 indicate?

the relationship is directly correlated

What does a correlation of 0 indicate?

the relationship is no correlation

What does a correlation of -1 indicate?

the relationship is inversely correlated

What can a correlation not tell us?

how much influence is exerted by the IV on the DV

Regression

quantifies the functional relationship where the IV is used to predict the DV

When do you use a linear regression?

when the DV is continuous

When do you use a logistic regression?

when the DV is dichotomous

Cox regression

used in survival analysis, DV is dichotomous

What is a covariate?

a continuous variable that is not the primary interest of the study but is known to affect the outcome

Why are covariates controlled for?

to see if the effect of the IV on the DV is influenced by it

How do confounding variables differ from covariates?

they are related to both the IV and DV, distorts results, a threat to validity

What are the various names for the IV?

x, predictor, exposure, factor, treatment

What are the various names for the DV?

y, response, outcome, disease, criterion

Simple linear regression

includes 1 IV and no covariates

Multiple linear regression

2 or more IVs, includes covariates

What does R squared represent?

the proportion of the variance in the DV explained by the IV's, ranging from 0-1 or 0-100%

What is the unstandardized regression coefficient B?

represents the amount of change in the DV due to change of 1 unit of the IV

What is the standardized regression coefficient?

represents which of the IV’s has the strongest effect on the DV; scales variables to standard deviation units

What are the types of review papers?

systematic, scoping and narrative reviews

Systematic reviews

identifies, appraises, and synthesizes all relevant studies on a specific topic to answer a targeted question

Scoping reviews

provide a descriptive summary of the sources of evidence without assessing the quality of the source

Narrative/literature review

summarizes studies within a topic but without a formal selection process

How are narrative and scoping reviews different?

Scoping reviews are conducted systematically and are reproducible while narrative reviews don’t have a formal selection process and are prone to selection bias

Meta-analysis

provides a quantitative summary by statistically combining results of the individual studies within the systematic review

Mixed methods systematic review

synthesis of findings from both qualitative and quantitative studies

What is the PICO method for designing a systematic review?

patient/population/problem, intervention, comparison, outcome

Main steps for designing a systematic review?

develop research problem and question, literature search, risk of bias assessment, evidence synthesis, discussion and conclusion

Random allocation concealment

implementing the random sequence in a way that conceals group placement from anyone who enters participants in a study

What is the benefit of including meta-analyses?

combines multiple study results, has greater statistical power, and detects treatment effects

Effect size

the magnitude of an effect

Mean difference (MD)

used when all studies in the meta-analysis measure the outcome on the same scale, shows the absolute difference between two groups

Standardized mean difference (SMD)

summary statistic in meta-analyses of continuous outcomes when the studies all assess the same outcome using a variety of scales; standardizes the results to a uniform scale

What is SMD often reported as?

Cohen’s d or d

As the SMD or Cohen’s d increases

the difference between the intervention and control groups increases

Small effect

d=0.2

Moderate effect

d = 0.5

Large effect

d = 0.8

When do we use MD and SMD?

when comparing studies with continuous outcomes (interval or ration data)

What is a dichotomous outcome?

either success or failure

When is RR and OR used?

when comparing studies with dichotomous outcomes

Inverse-variance weighting

using standard error to calculate the weight for each study

Heterogeneity

the variation in the direction and magnitude of the effect size for the studies

What is the line of no effect?

the vertical line in the graph; represents the null hypothesis

What does the horizontal line in forest plot graphs represent?

the 95% confidence interval

What is the diamond on the graph in forest plots?

the summary estimate, if it crosses the line of no effect, it isn’t statistically significant

Which term refers to a variable that is related to both exposure and outcome and can bias an observed association if not controlled?

confounding variable

Recall bias is a limitation in which type of epidemiological study?

case-control study

If a dependent variable is continuous, _______  is the correct type of regression analysis to use.

linear regression

What does inverse-variance weighting in meta-analysis give more weight to?

individual studies with larger sample sizes

Which type of review is most susceptible to selection bias due to lack of a systematic process?

narrative review

What is the primary purpose of a scoping review in nutrition research?

To map the existing literature and identify gaps for future research

Which type of review is primarily designed to answer a specific clinical question by identifying, appraising, and synthesizing all relevant studies using a pre-defined protocol?

systematic review

What does a forest plot typically illustrate in a meta-analysis?

Individual study effect estimates and a pooled summary estimate

In meta-analysis, what is the primary purpose of calculating an effect size?

To quantify the magnitude of the intervention’s effect across studies