Study Designs Flashcards
“research”
a rigorous and systematic investigation of any phenomenon
rigorous
refers to quality of research - using appropriate methods and analytical techniques
systematic
refers to the practice of conducting a research in a logical and orderly, disciplined manner
Types of Study designs (8)
- ecological/correlation
- cross-sectional survey
- cohort study
- case-control study
- RCT
- clinical trial
- animal study
- systematic review or meta-analysis
Descriptive/Exploratory Designs
- collected to ID potential associations
- useful to make recommendations for future research
- collect baseline data
- useful in developing hypothesis
Nutritional Epidemiology
the study of the effect of nutrition on disease
⇒ ecological, case control, cohert
Ecological/Correlation
- observation of a group of people usually within 1 geographical region and loo at diet with some outcome
- or compare a factor of interest among different groups and the estimates of the intake of food/nutrients for those groups
Correlation strengths
- used as an exploratory tool
- yielded considerable knowledge about dietary factors and chronic diseases
- used to generate hypothesis
Hard to establish causality
- most chronic diseases are multifactoral
- chronic diseases have long latency periods/ when do they start
- hard to accurately estimate intake
- multiple confounding factors
Cross sectional Surveys
- a snap shot of an entity
- association between an indicator of nutrition and a health characteristic measured at the same time
strengths of Cross sectional surveys
- snap shot of entire region
- used to ID factors that may affect the level of risk factor
- generates hypothesis
Case Studies
- collect info on one or a few individuals
- collect comprehensive information about a cluster of characteristics
- usually do not manipulate the environment
- findings are very specific to the group you selected
Surveys
- used to collect information on a larger group of individuals
- can be researcher or self-administered
- cross sectional surveys collected info at one point in time
longitudinal surveys look at the issue over time
Analytical Designs
exploratory/epidemiological ⇒ case control studies, cohort studies (next)
Explanatory - test ⇒ clinical trials, animal studies
Case Control Studies (retrospective)
- a group of subjects with the disease is compared with a group without the disease and their exposure history is compared
- collect retrospective info on diet and lifestyle
Strengths of Case Control Studies
- very useful for less frequent diseases
- cost and time efficient
- investigate hypothesis of causal relationship
- used to explore aetiology - very useful for strong relationships
Cohert Study (prospective)
- a population of people are identified; measure baseline and then look at disease occurrence overtime
- people are followed over time, having a common characteristic or factors of interest
Examples of Cohert Studies
Framingham studies Nurses Health Study EPIC Kadoorie Study APrON
Strengths of Cohert Studies
- evaluation of time-related associations
- baseline info helps decrease recall bias
- examine possible causal factors exposures
Cohert Studies (retrospective)
- looks backwards in time
- the outcome has occurred
- go back and collect data - can only include what has previously been recorded
- collect what might have contributed to the outcome
Criteria for Distinguishing Association from Causation
Temorality ⇒ cause precedes effect
strength of association ⇒ large risk
dose-response ⇒ greater exposure - higher frequency of effect
consistency ⇒ repeatedly observed by independent investigators
biological plausibility ⇒ causal interpretation should fit with current science
Randomized Clinical Trial
- subjects randomly assigned to diets and the development of the disease/risk factor for a disease is measured
Strengths of RCT
- powerful design for evaluation practice and medical/dietary treatments
- proves feasibility and safety of treatments, compare one or more treatments
Limitations of RCT
- Compliance
- can’t blind dietary or lifestyle factors
- what is the interpretation of negative findings
- problems with long term intervention
Steps in Clinical Trial
- hypothesis
- study designs and duration
- subject selection ⇒ inclusion/exclusion, recruitment, screening
- diet
- study procedures ⇒ sample size, randomization, monitoring, compliance
- statistics
animal studies
critical to understanding mechanisms of cause and effect relationship
proof to rationalize human trial
biological plausibility
PEN - practical based evidence nutrition
- series of knowledge pathways, each focussing on a topic
- developed from key practice questions and evidence-based answers
Systematic reviews
keep up with evidence, resolve uncertainty of contradictory results, helps demonstrate lack of evidence, future research?
- subjects are papers, not people
Steps in a systematic review (8)
- formulate a question
- choose a database
- search/limitations
- eligibility criteria
- quality assessment (risk of bias)
- data extraction - directly as is
- synthesis of data findings
- sensitivity analysis - which quality results are included?
Types of Bias in Systematic Reviews (6)
selection bias (what was their reason for entering, sampling bias)
outcome reporting bias (some outcomes not reported)
citation bias (using one papers references for other reviews)
publication bias (published only positive results)
language bias
duplicate bias