NUTR 400 Quiz 2
What are ecological studies?
- Observe variable in different geographical regions (such as countries) -> Describe and compare the different populations -> Establish correlations between variables - Used to generate rationale for additional studies
Steps of a Systematic Review 4) Data Collection What is required on the flow diagram for the PRISMA Reporting Guidelines? (showing the number of records that were ... 5)
- PRISMA -> Preferred Reporting Items for Systematic Reviews and Meta Analysis - Requires a flow diagram showing the number of records that were: -> Identified -> Removed as duplicates -> Excluded at title/abstract screening -> Excluded at full-text screening, with reasons for exclusion -> Included in analysis - Example #1: Burckhardt et al. (2016), Figure 2, p. 15 (you just write n = # studies at each step)
Experimental Studies What are cross-over studies?
- Participants randomized to start with intervention or placebo - After a followed wash-out period, receive the other arm - Each participant serves as its own control; all participants receive intervention and placebo ___________________________ 2 Groups - intervention 1 month -> wash out period -> placebo 1 month - placebo 1 month -> wash out period -> intervention 1 month
Steps of a Systematic Review 1) Defining the Research Question - What are the research questions from the assigned readings? Burckhardt et al. (2016) What is "the efficacy and safety of omega-3 polyunsaturated fatty acid (PUFA) supplementation for the treatment of people with dementia"?
- Population = people with dementia - Intervention = omega-3 PUFA supplementation - Comparison = control (no omega-3) - Outcome = dementia (look in methods section)
Steps of a Systematic Review 1) Defining the Research Question - What are the research questions from the assigned readings? Shoenakeret al. (2016) What is the "current evidence from observational studies among reproductive-aged women on the association of energy, nutrients, foods, and overall dietary patterns with the development of GDM"?
- Population = reproductive-aged women - Exposure = energy, nutrients, foods, and overall dietary patterns - Outcome = development of GDM - Study Design = observational studies
Steps of a Systematic Review 4) Data Collection What are the PRISMA Reporting Guidelines?
- Preferred Reporting Items for Systematic Reviews and Meta Analysis - Requires a flow diagram
Steps of a Systematic Review 1) Defining the Research Question The question must be focused and clear. What are the PICO(TS) elements?
- Question must be focused and clear - Addresses PICO(TS) elements -> Population -> Intervention or Exposure -> Comparison -> Outcome -> Time -> Study design - Informs the inclusion/exclusion criteria _______________________________________________ - PICO is one of the methods used to develop a focused and clear research question - Specific inclusion and exclusion criteria are based on PICO(TS) elements -> Determines which studies are eligible for inclusion in the SR
Why is there a need for SRs (3)?
- Rate of scientific publications is increasing astronomically - Difficult to stay up to date with all the new emerging evidence - One stop shop ________________________________________________ - There is still no single repository reliably showing the true number of randomized trials
Experimental Studies What is the difference between these 3 types of variables?
1) Independent = Intervention/treatment - Can be >1 independent variable (e.g. factorial design) 2) Dependent = Outcome(s) 3) Confounding - Variables that can affect the dependent variable and/or the response to the independent variable ________________________________________ Example - Independent variable = level of dietary protein - Dependent variable = change in body weight - Confounding variables? (exercise, kcal intake, protein source, etc.)
What is the name for qualitative and quantitative variables? What are the types in each? Quantitative (2) Qualitative (3)
1) Numerical A) Discrete B) Continuous (e.g. blood glucose) 2) Categorical A) Binary (e.g. yes/no) B) Ordinal (e.g. very hot, hot, warm, cold, very cold) C) nominal (e.g. what country were you born in? - and choose from a list)
What are 4 examples of non-random sampling methods?
1) Quota sampling - E.g. know how many males/females you want, how many people of a certain income bracket you want, etc. - You accept people into the study based on the quotas you have established 2) Convenience/accidental sampling - E.g. put up posters for your research and people come in 3) Purposive/expert sampling 4) Snowball sampling - People who participate in a study recommend the study to one of their friends/relatives
What is bias?
= A systematic deviation from the truth
Observational Studies What is another name for observational studies?
= Also called non-experimental because there is no intervention - Are used in descriptive, correlational and explanatory research
What is data dredging/fishing?
= Analyzing large volumes of evidence to find statistically significant results without first specifying a hypothesis which would explain causality
Steps of a Systematic Review 2) Literature Search - What are the characteristics of a systematic search? (6)
= CORD CJ 1) Clear 2) Orderly 3) Reproducible 4) Documented 5) Consistent 6) Justified - Best to consult with a research librarian for SR searches
What is the risk of using a small sample size?
= Failing to find significant differences in sample even if significant differences exist in the larger population - If a study does find significance with a small sample size, then that means that the sample was large enough
What is the 'cohort' in prospective cohort studies?
= Individuals with a common characteristic
Experimental Studies What is the gold standard for experimental study designs?
= Randomized Double-Blinded Controlled Trials (RCT) - Expensive to do and sometimes not feasible
What is sample size for qualitative studies usually based on?
= Saturation point
What is the main purpose of the qualitative mode of inquiry?
= To describe variation in something - Individual experiences, meanings, perceptions, feelings - More flexible study designs - Identify + describe themes/concepts
Experimental Studies Can there be more than one independent variable in a study?
= yes - 2x2 factorial design -> 2 two-level treatments, 4 groups __________________________________________________ Example - Cystine supplementation or no cystine - Ibuprofen or no ibuprofen Groups - cystine + ibuprofen - ibuprofen alone - cystine alone - controls
Steps of a Systematic Review 1) Defining the Research Question - What is the template for PICO(TS) questions? Examples -> Intervention Study -> Diagnostic Test -> Etiology and Risk Factors (e.g. observational studies)
Intervention - In preterm infants (P), how does hydrolyzed formula (I) compare to standard cow's milk-based formula (C) affect the incidence of necrotizing enterocolitis (O) within the initial hospital admission (T)? Diagnostic Test - In men over 60-years-old (P), is serum homocysteine (I) compared to HDL:LDL (C) more accurate in diagnosing coronary heart disease (O)? Etiology and Risk Factors - Are children (P) who have a parent that smokes (I) compared to those with non-smoking parents (C) at increased risk for obesity (O) between the ages of 20-30 years-old (T)?
Steps of a Systematic Review 1) Defining the Research Question - What are the templates for PICO(TS) questions? -> Intervention Study -> Diagnostic Test -> Etiology and Risk Factors (e.g. observational studies)
Intervention -> In ____(P), how does ____(I) compared to ____(C) affect ____(O) within ____(T) Diagnostic test - In ____(P), is ____(I) compared to ____(C) more accurate in diagnosing ____(O)? Etiology and risk factors - Are ____(P), who have ____(I) compared to those without ____(C) at ____ risk for ____(O) over ____(T)?
What is the difference between low and high risk of bias (RoB)?
Low RoB = Methodologically sound, more likely that results reflect the 'true' effect High RoB = Results may not be 'true'; inappropriate methods may have influenced the findings - Methods to assess risk of bias depend on what study designs are being collected (randomized, no randomized, mixed methods) ________________________________________ Risk of Bias for RCTs - RoBtool has set domains (see graph below) -> Included studies are ranked as Low, Unclear or High RoB for each domain _________________________________________ - Skipped this
What are experimental study designs?
- Involves an intervention and tests a hypothesis -> Cause and effect, mechanisms - Also called intervention studies - Studies are controlled (to a varying degree) to minimize impact of other factors - Easier in animal studies, more difficult in human studies
Steps of a Systematic Review 5) Quality Assessment (QA) - Newcastle Ottawa Scale for observational studies
- 3 domains: selection of study groups, comparability of study groups, and ascertainment of outcomes - "Star system": stars awarded for high quality methods within each domain -> e.g. Shoenakeret al. (2016) Supplementary Tables S2 and S3. - Overall assessments of how well the included studies addressed each domain of the scale should be discussed in the Results and Discussion _________________________________________ - Skipped this
What are cross-sectional studies?
- A 'snap-shot' description of a group or groups via measurements at a single point in time -> E.g. surveys, dietary intake assessment, anthropometrics, level of nutrients in blood - Used to describe the population and can be used to examine correlations between variables ___________________________ - each subject only comes in once
What is a systematic review (SR)?
- A synthesis of evidence answering a specific research question Methodology of SR - Structured process to identify, collect, synthesize, and appraise evidence - Designed to ensure transparency, accuracy in data collection and analysis, and avoid bias in results and conclusions - Team effort (context expert, systematic review methods expert, statistical expert, information specialist)
Meta-Analysis - What is a forest plot?
- An essential tool to summarize info on individual studies - Gives a visual suggestion of the amount of study heterogeneity, and show the estimated common effect, all in one figure _____________________________________ https://www.students4bestevidence.net/blog/2016/07/11/tutorial-read-forest-plot/ 95% CI = range of values within which you can be 95% certain the true value lies - If the horizontal line crosses the line of null effect what that is effectively saying is that the null value lies within your CI and hence could be the true value - Any study line which crosses the line of null effect does not illustrate a statistically significant result - The diamond represents the point estimate and confidence intervals when you combine and average all the individual studies together - This combined value effectively groups all the participants from all the individual studies - The CI range for this result should be the smallest on the forest plot - The rules about crossing the line of null effect are still true here: if the horizontal tips of the diamond cross the vertical line, the combined result is potentially not statistically significant
Steps of a Systematic Review 6) Data Synthesis and Analysis
- Analysis may be descriptive, quantitative, or both - An important step for quantitative data analysis is determining whether or not data from included studies should be pooled -> Only similar (homogeneous) data should be pooled) Heterogeneity = "variability among studies" -> Clinical differences (participants, interventions, outcomes) -> Methodological differences (study design and risk of bias)
Lecture Systematic Reviews and Meta-Analysis Dr. Sarah Elliot Dr Rhonda Bell
- Both assigned readings for today will not have questions in the quiz on them ever - But they are perfect examples of systematic reviews
Experimental Studies What are repeated measure studies?
- Involves multiple measures of the same dependent variable taken on the same subjects (wiki) - Can establish when the intervention becomes effective
Steps of a Systematic Review 2) Literature Search To avoid publication bias, SR researchers should search what? (2)
- Multiple databases (MEDLINE/PubMed, EMBASE, CINAHL, Web of Science, PsycINFO, ERIC, etc.) - Grey literature
Steps of a Systematic Review 4) Data Collection - What are some of the important data to collect? (3)
- Characteristics of included studies (e.g. study design, dates, PICO elements, n in each comparison group) - Outcomes (quantitative, qualitative or both? Dichotomous, continuous, ordinal, time-to-event?) - Risk of bias assessment (see quality assessment section) _________________________________________ - Be prepared that not all the data from included studies will fit neatly into the DE form Be Aware Of - Units (e.g. VitD reported in mcg or IU?) - Time points that outcomes were measured at (e.g. 1 vs. 6 months) - Scales, equipment, or method of measuring data - Is this comparable between studies? ______________________________________________ - Usually, Table 1 of an SR describes characteristics of included studies -> Should address all PICO elements - Examples: Burckhardt et al. (2016), p. 38; Shoenakeret al. (2016), supplementary data file
What are prospective cohort (longitudinal) studies?
- Collect data in the present and follow-up in the future - Examine if factor (exposure) increases or decreases risk of health outcomes - Can also examine progression of condition or disease; changes with aging or maturation; can use data for nested case-control studies Examples - Framingham Heart Study - Nurses Health Study - EPIC - Tomorrow Project - APrON
What are case-control studies?
- Compares two groups -> Individuals who have a condition/disease and those without it (controls) - Look at variables retrospectively (such as risk factors) -> Retrospective interviews, surveys/questionnaires, records
Steps of a Systematic Review 2) Literature Search
- Create search "concepts" from the research question - Databases usually differ with respect to format, use of special characters, etc -> These nuances must be translated as best as possible between databases to ensure consistent searching - Should have the search strategy documented and available upon request -> MEDLINE search strategy is often reported as an appendix - Other database searching features -> Filters (e.g. RCT filter, pediatric filter) -> Limits (e.g. English only, publication dates) - Ensure the decision to use these options are justified (don't just google!) _______________________________________________ - E.g. In preterm infants, how does hydrolyzed formula compared to cow's milk-based formula affect the incidence of necrotizing enterocolitis (NEC) within the initial hospital admission? Concept (1=NEC -> 2=formula -> 3=infant) - necrotizing enterocolitis AND artificial formula AND premature infant OR - intestinal diseases AND infant formula AND low birth weight OR - Gastro* or gut AND hydrosyl* or hydroly* or intact protein AND neonatology
What are some of the ways to do reviews?
- Critical Literature Review - Mapping Review - Systematic Literature Review - Rapid Review - Scoping Review - Overview
Learning Objective
- Define a systematic review - Understand the difference between a systematic review and narrative review or critical literature review - Know the steps of a systematic review and rationale for each step - Recognize the differences between a systematic review of randomized controlled trials and observational studies - Understand what a meta-analysis is and how to interpret a forest plot
Types of Research Based on Objectives - What is descriptive research?
- Describes a group of individuals -> May or may not have a comparison group Examples: - Prevalence of type 2 diabetes in Alberta - Attitudes of students towards healthy eating - Characteristics of individuals seeking dietetic counselling - Insulin-sensitivity in adolescents - Nutrition needs of a community
Match the following study description to its type: - Effect of a prenatal nutrition program on infant birthweight - Prevalence of T2D in Alberta - Body weight and screen time - Descriptive - Correlational - Explanatory
- Descriptive - Prevalence of T2D in Alberta - Correlational - Body weight and screen time - Explanatory - Effect of a prenatal nutrition program on infant birthweight
What is the purpose of correlational research?
- Establish or explore relationships Examples: - Body weight and screen time - Fibre intake and colon cancer - Ghrelin concentration and appetite
Experimental Studies What are before and after (pre-post) studies?
- Examines whether participants in an intervention improve or regress during the course of the intervention - Baseline measure of dependent variable -> Start intervention (independent variable) -> Measure dependent variable again post-intervention - Did the independent variable change the dependent variables?
What is the purpose of explanatory research?
- Explains why a relationship exists; evidence to support causality Examples: - Effect of prenatal nutrition program on infant birthweight - Association between folate intake and risk of NTDs - Impact of fatty acids on gene expression
What type of studies are experimental study designs? (based on objectives)
- Explanatory studies, but can also have descriptive and correlational aspects
What is grey literature?
- Information not produced through commercial publishing Examples -> Clinical trials registries -> Conference proceedings -> Dissertations and theses -> Government or food and drug company reports -> Google and social media (newsletters, blogs)
Why are SRs important? (6)
- Reduces time and effort spent on researching a topic by combining all relevant data into one study - Identifies what has already been done in the field -> Ensures research resources are used efficiently -> Highlights gaps in literature to focus future efforts ("empty reviews") - Advances research (by assessing the quality of the evidence) - Increases statistical power by pooling similar samples (↑N) - Foundation for clinical practice guidelines - Improves decisions __________________________________________ - Healthcare providers, consumers, researchers, and policy makers are inundated with unmanageable amounts of information, including evidence from healthcare research. It is unlikely that all will have the time, skills and resources to find, appraise and interpret this evidence and to incorporate it into healthcare decisions. Cochrane reviews respond to this challenge by identifying, appraising and synthesizing research-based evidence and presenting it in an accessible format
Steps of a Systematic Review 4) Data Collection What happens after the PRIMSA analysis?
- Review team creates a data extraction (DE) form that all reviewers use (a standard form) - Outcomes and data analysis plan should be determined a priori -> Avoids "data dredging/fishing" -> Avoid bias in reporting data (primary vs. secondary outcomes) _______________________________________ - Not that important
Steps of a Systematic Review 3) Study Selection (Screening)
- Review team creates and pilots an Inclusion/Exclusion Criteria form - 2 independent reviewers to screen studies for eligibility - 2 phases of screening
Steps of a Systematic Review 5) Quality Assessment (QA) - What is quality assessment?
- Reviewer assesses how well studies were conducted -> Helps determine reliability of results - Helps the reader determine how included studies were conducted and reliability of their results ____________________________________________ - 2 independent reviewers should complete QAs - QA findings should be integrated into the results and conclusions
Summary
- SRs follow a rigorous process to identify, collect, synthesize and critically appraise the evidence available on a defined research question -> Rigor is necessary to avoid bias - If done well, SRs provide the highest level of evidence for practice and policy decision-making
What is sampling?
- Selecting a few from a larger population to predict parameters of the larger population -> Larger population (N) -> Smaller sample population (n)
What is a meta-analysis?
- Statistical approach to combine quantitative results (I.e. pooling of studies from the SR in the data synthesis and analysis step) - Produces a "point estimate" to quantify the overall effect of a comparison - Typically presented as a forest plot _______________________________________________________ - Only similar (homogeneous) data should be pooled) - Heterogeneity = variability among studies
What is the methodology of a systematic review (SR)?
- Structured process to identify, collect, synthesize, and appraise evidence - Designed to ensure transparency, accuracy in data collection and analysis, and avoid bias in results and conclusions - Team effort -> Context expert, systematic review methods expert, statistical expert, information specialist ___________________________________________ - Not that important
What are some examples of the mixed methods mode of inquiry?
- Studies can use both quantitative and qualitative methods (mixed methods) - Quantify and describe Example: -> Quantitative method to assess beliefs about healthy eating using scale-based survey -> Qualitative method to explore why individuals have certain beliefs around healthy eating using interviews
Steps of a Systematic Review 7) Discussion Section of an SR - What is included?
- Summarize main findings and strength of evidence - Discuss limitations regarding certain studies, outcomes (RoB) or the review as a whole - Strengths of the review - Interpret results, considering relevance to key stakeholders and implications for future research - Implications for current practice
What are the limitations of SR? (2)
- Synthesis is only as good as the data put into it - Data analysis limited by available body of literature (E.g. if there are no studies on a topic)
Which of these study designs provides the highest level of evidence for causality? (rank them) RCTs Ecological Studies Case control studies Cross sectional studies Cohort studies (prospective and retrospective)
- Systematic reviews = best 1) RCTs 2) Prospective cohort studies 3) Retrospective cohort studies 4) Case control 5+6) Ecological and Cross sectional
What is power (of an analysis)?
- The chance of detecting a real effect as statistically significant if it exists - Many individual studies are too small to detect small effects -> But when several are combined there is a higher chance of detecting an effect
Why is sample size important?
- The higher the variation of the parameter being assessed, the larger the sample size needed to accurately estimate
What are retrospective cohort studies?
- The same as prospective, except the data is from the past (e.g. medical charts or other records; based on recall) - Also referred to as historic cohort studies
Steps of a Systematic Review 4) Data Collection - What else might you want to collect?
- Total study duration - Setting - Diagnostic criteria - Age, sex, ethnicity, SES, comorbidities - Intervention details (dose, frequency, method of measurement, etc.) - Outcome scales (upper and lower limits, and whether high or low score is good) - Estimate of effect with CI, p-value - Funding score - Key conclusions - References to relevant studies - RoB components (sequence generation, allocation concealment, outcomes and time points collected and reported, missing participants)
When should you use non-random/non-probability sampling?
- When you can not identify everyone in the larger population or is not feasible to randomly choose from larger population
How can the types of research commonly used in nutrition be categorized? (2)
1) Based on mode of inquiry 2) Based on objectives - Type of research and study design does have an impact on what questions can be answered -> Objectives and conclusions must be appropriate for the type and design
What are the 3 types of studies in each category? 1) Based on mode of inquiry 2) Based on objectives
1) Based on mode of inquiry -> Quantitative, qualitative, mixed methods 2) Based on objectives -> Descriptive, correlational, explanatory
What are the steps in conducting a systematic review? (7)
1) Clearly define the research question 2) Comprehensive literature search 3) Study selection 4) Data collection 5) Quality assessment (QA) 6) Data synthesis and analysis 7) Discussion
Observational Studies Provide 5 examples of observational study designs
1) Ecological studies 2) Cross-sectional studies 3) Case-control studies Cohort studies 4) Prospective 5) Retrospective
What are the 2 broad categories of study designs?
1) Experimental 2) Observational
Steps of a Systematic Review 6) Data Synthesis and Analysis What are the advantages of MAs? (5)
1) Increases the power of analysis 2) Improves precision of effect estimate (narrower confidence interval) 3) Demonstrates if consistency of an effect is maintained when multiple studies are combined (individual studies are not likely to have the exact same population or intervention) -> Or provides reasons for differences in effect estimates 4) Resolves discrepancies between individual studies -> Point estimate of pooled data closer to the "true" effect 5) Generates new hypotheses
Experimental Studies What are the 3 main types of variables in experimental study designs?
1) Independent 2) Dependent 3) Confounding
Experimental Studies What are the components of gold standard experimental design? (5)
1) Randomized - Participants randomized to receive the treatment or not receive the treatment by chance 2) Double Blind - Investigators and participants do not know which group assigned to each participant 3) Controlled - Confounding variables controlled and/or measured 4) Well defined participant selection 5) Designed to answer one central question (with one primary outcome) _______________________________________________________ - Experimental studies should add as many of these as possible - Strengths and limitations of experimental study designs depends on the degree to which the gold standard components are incorporated - Additional aspects of experimental study design that have strengths and limitations: -> What outcome assessed and methods used to assess -> Intervention and placebo composition -> Parallel and cross-over design -> When outcome assessed
What are the 2 sampling strategies?
1) Randomized/probability sampling - Everyone in larger population has equal probability of being selected 2) Non-random/non-probability sampling
What is the difference between randomized and non-randomized sampling?
1) Randomized/probability sampling - Everyone in the larger population has equal probability of being selected in the sample - Basic methods -> advanced computer programs 2) Non-random/non-probability sampling
What is the hierarchy of evidence for human research? Rank the following studies Cohort studies SRs and MAs Cross-sectional studies RCTs Cohort studies Editorials, expert opinions, case reports; animal and cell based studies
1) SRs and MAs 2) RCTs 3) Cohort studies 4) Case-control studies 5) Cross-sectional studies 6) Editorials, expert opinions, case reports; animal and cell-based studies - As you move up the pyramid the study designs are more rigorous and allow for less bias or systematic error that may distract you from the truth ____________________________________________ - In nutrition we often are not able to conduct RCTs -> SRs + RCTs very good for certain types of questions -> But not all questions can be answered by SRs or RCTs
What are 3 methods of randomized/probability sampling?
1) Simple random sampling - Entire population 2) Stratified random sampling - Divides the population into smaller groups/strata based on shared characteristics and selects from them 3) Cluster sampling - Total population divided into clusters and a simple random sample of the groups is selected - Sampling is done on a population of clusters ______________________________________________ - Definitions are from wiki
The sample size (quantitative) needed is based on? (3)
1) The level of confidence desired 2) The degree of accuracy you want (in the estimate of the primary outcome variable) 3) The estimated level of variation in the primary outcome variable
Steps of a Systematic Review 3) Study Selection (Screening) What are the 2 phases of screening?
1) Title and abstract 2) Full text -> Identify and record reasons for exclusions (e.g. wrong study design, wrong population, wrong exposure, wrong comparison, wrong outcome, duplicate, etc.) ________________________________________ - Review team creates and pilots an Inclusion/Exclusion Criteria form - Need at least 2 independent reviewers to screen studies for eligibility - Use the same I/E form for full-text screening
Class discussion of Burckhardt et al. (2016) meta-analysis
1) Which treatments are being compared in this meta-analysis? 2) Which outcome is the meta-analysis looking at? 3) How many primary studies were included in this analysis? 4) What is the total pooled n for each comparison group? 5) Which primary study holds the most weight in the analysis? 6) Which statistic is being used for the overall point estimate? 7) What is the overall point estimate? 8) What is the confidence interval of the point estimate? Is this wide or narrow? 9) What is the statistical heterogeneity (%)? Is it appropriate to combine studies? __________________________________________ - skipped
What are the benefits (2-3) and limitation (1) of retrospective when compared to prospective cohort studies?
Benefit = Less time consuming and expensive = No conditioning or reactive effects (no attrition?) Limitation = Data desired may not be present (not planned at the time of data recording)
How do case control and retrospective cohort studies differ?
Case Control - Retrospective - 2 identical groups except for outcome/disease - Start from outcomes and look back to see if there is a SS difference in rate of exposure - Odds ratio (is main outcome measure) Retrospective Cohort - Retrospective - Already know exposure and outcome in the cohort - Study association of exposure to disease outcome - Exposures are defined before looking at the existing outcome data to see if exposure to a risk factor is associated with a SS difference in outcome development rate - Identical groups except for exposure status ______________________________________ Case Control - Objective: to determine the association between mother's dietary diversity and stunting. The study was conducted at a hospital in Dhaka, Bangladesh from November 2016 to February 2017. Data were collected from mothers of stunted children less than 2 years and non-stunted children less than 2 years. Mothers were asked to recall their consumption of 10 defined food prior to the interview as per Guidelines for Minimum Dietary Diversity for Women Retrospective Cohort - Objective: to assess whether in utero exposure to gestational diabetes mellitus increases the risk for long-term endocrine morbidity of the offspring. The study included all singleton born to women who delivered between 1988 and 2014in a tertiary medical center. Data were collected from the computerized perinatal database of the obstetrics and gynecology department and the computerized hospitalization database
How do descriptive, correlational, and explanatory research differ?
Descriptive - Describes a group of individuals - May or may not have a comparison group - E.g. Prevalence of type 2 diabetes in Alberta Correlational - Establish or explore relationships - E.g. Body weight and screen time Explanatory - Explains why a relationship exists; evidence to support causality - E.g. Effect of prenatal nutrition program on infant birthweight
Which types of studies (based on objectives) are experimental and observational studies usually?
Experimental - Explanatory studies - But can also have descriptive and correlational aspects Observational - Used in descriptive, correlational, and explanatory research
How do experimental and observational studies differ?
Experimental (intervention studies) - Involves an intervention and tests a hypothesis (cause and effect, mechanisms) - Studies are controlled (to a varying degree) to minimize impact of other factors Observational (non-experimental) - There is no intervention
What is the difference in the main purpose of the quantitative and qualitative mode of inquiry?
Quantitative = To quantify the extent of variation in something - Structured with pre-determined study designs - Statistical analysis to interpret data Qualitative = To describe variation in something - Individual experiences, meanings, perceptions, feelings - More flexible study designs - Identify + describe themes/concepts ________________________________________________________________ Mixed Methods - Studies can use both quantitative and qualitative methods Example: -> Quantitative method to assess beliefs about healthy eating using scale-based survey -> Qualitative method to explore why individuals have certain beliefs around healthy eating using interviews
What are the strengths (2) and limitations (2) of ecological studies?
Strengths 1) Can explore interesting associations with more variation between geographical locations 2) Supports further research Limitations 1) Other factors likely different between geographical locations 2) Data often on population rather than individual level (e.g. food disappearance data)
What are the strengths (4) and limitations (2) of cross-sectional studies?
Strengths 1) Can get detailed description of individuals in a group at one point in time; multiple variables concurrently measured 2) Can compare between two or more groups 3) Identifies associations between variables 4) Generates hypotheses Limitations 1) Only one point in time (can not establish sequence of events -> prevents ability to assess causal associations) 2) Sample selection important for generalizations
What are the strengths (3) and limitations (4) of prospective cohort studies?
Strengths 1) Data collection is planned; recall bias is minimized 2) Can establish incidence of disease/condition; identify sequence of events 3) Assess causal associations (risk ratio, rate ratio) Limitations 1) Attrition is likely 2) Conditioning and reactive effects if repeated measurements 3) Lengthy (can be a number of years long) 4) Expensive causal inference but not direct cause-effect
What are the strengths (3) and limitations (3) of case-control studies?
Strengths 1) Examine less frequent conditions/disease 2) Can explore etiology; assess causal associations (odds ratio) 3) Cost and time efficient Limitations 1) Data may not be available and/or relies on recall 2) The right controls needed (matching) 3) Causal inference but not direct cause-effect
What are the strengths (4) and limitations (2) of retrospective cohort studies?
Strengths 1) Less expensive and time consuming than prospective 2) No conditioning or reactive effects 3) Can establish incidence of disease/condition and identify sequence of events 4) Assess causal associations (risk ratio, rate ratio) Limitations 1) Data may not be available and/or relies on recall 2) Causal inference but not direct cause-effect
What is the difference between systematic and random error?
Systematic Error - Multiple replications of the same study would reach the wrong answer on average Random Error (imprecision) - Multiple replications of the same study will produce different effect estimates b/c of sampling variation even if they would give the right answer on average
How does the value of a SR and a CLR differ?
Systematic Review - Comprehensive, systematic summary of high quality evidence - Support for evidence based practice Critical Literature Review - Summary and understanding of a topic
What are the differences between a systematic literature review and a critical literature review? 1) Research Question 2) Literature Search 3) Selection of Included Studies 4) Data Analysis 5) Quality Assessment 6) Number of Authors 7) Value
Systematic Review 1) Research Question = clearly defined 2) Literature Search = comprehensive, multiple databases, grey literature 3) Selection of Included Studies = pre-specified criteria 4) Data Analysis = quantitative, descriptive, or both 5) Quality Assessment = required 6) Number of Authors = usually ≥3 7) Value = comprehensive, systematic summary of high quality evidence; support for evidence based practice Critical Literature Review 1) Research Question = general or specific 2) Literature Search = subjective approach 3) Selection of Included Studies = subjective approach 4) Data Analysis = mostly descriptive 5) Quality Assessment = usually not included 6) Number of Authors = ≥1 7) Value = summary and understanding of a topic
What are the differences between a systematic literature review and a critical literature review? 1) Research Question 2) Literature Search 3) Selection of Included Studies 4) Data Analysis 5) Quality Assessment 6) Number of Authors
Systematic Review (SR) 1) Research Question = clearly defined 2) Literature Search = comprehensive, multiple databases, grey literature 3) Selection of Included Studies = pre-specified criteria 4) Data Analysis = quantitative, descriptive, or both 5) Quality Assessment = required 6) Number of Authors = usually ≥3 Critical Literature Review (CLR) 1) Research Question = general or specific 2) Literature Search = subjective approach 3) Selection of Included Studies = subjective approach 4) Data Analysis = mostly descriptive 5) Quality Assessment = usually not included 6) Number of Authors = ≥1
