worksheet C
preiou5
study_pyramid_design.pptStudy Design, Levels of Evidence, and Relationship to PICO
Management of Evidence-Based Dental Hygiene Care
DHY 8330
*
Lecture Objectives
At the end of this presentation the dental hygiene student will be able to:
- Distinguish between primary and secondary research
- Identify characteristics of several types of study designs
- Identify levels of evidence
- Correlate study design to the classifications of PICO questions
*
Scientific Evidence
Well-designed and well-controlled research investigations minimize sources of bias. Usually one research study does not constitute “the evidence.”
Patient Preferences
or Values
Clinical/Patient Circumstances
Experience and
Judgment
Scientific Evidence
*
What are we looking for and Why?
Challenge when only one study found:
Cautious: later it could be contradicted, or
found to have used poor methods or
test only efficacy
In vitro Research
- A study carried out in the laboratory setting with no direct involvement of living human subjects
- This study might involve human tissue or blood, serum, urine, gingival crevicular fluid, saliva
- This study might use cultured human cells
*
Animal Research
- A study carried out in the laboratory setting with no direct involvement of living human subjects
- This study involves animals as models of human disease
*
Expert Opinion
- Individual opinions based on experience or partial assessment of evidence
*
Survey
- A study performed for the purpose of collecting general or specific information on particular subjects from a target group or population
- Information is generally not collected for purposes of hypothesis testing, but hypothesis generation
*
Cross-Sectional Study
- Studies performed over a limited, short, specific period of time (i.e., seconds, minutes or days)
- Both the independent (exposure, risk factor, experimental) and dependent (outcome or response) variables are measured at the same time
*
Case Report and Case Series
- A case report is a publication on the treatment of a single patient
- A case series is a publication on the treatment of several patients or a collection or reports
- These are not statistically valid, no control
- May lead to hypothesis-based research
*
Retrospective: subjects already have disease compared with group without disease
Problem: looking back in time
rely on subjects recall
Prone to recall bias and extraneous variables
Case-Control Studies
- Studies possible associations between a disease and a risk factor
- Compare subjects who have an identified disease or condition with those who do not
- Case-control studies are retrospective and rely on subjects recall (recall bias)
*
All these factors contribute to reducing bias and increase validity
Prospective
See “Glossary” on Blackboard site
Lung Cancer
No Lung Cancer
No Exposure
Exposure
To tobacco
Exposure
To tobacco
No Exposure
Back in time
Start here
Case-Control Studies are retrospective
- At the beginning of the study, subjects who have disease and who don’t have disease are identified and grouped
- Information about past behaviors or exposures for each group is collected (patient records, interviews)
- Common behaviors or exposures that might have contributed to the development of the disease are identified
*
Retrospective: subjects already have disease compared with group without disease
Problem: looking back in time
rely on subjects recall
Prone to recall bias and extraneous variables
Cohort Studies
- Try to find an association between an exposure or risk factor and the development of a disease or condition
- Two groups – one group has the risk factor or exposure, and one group does not have the risk factor or exposure
- The two groups are followed for some time and the development of the disease or condition is compared between the two groups
*
All these factors contribute to reducing bias and increase validity
Prospective
See “Glossary” on Blackboard site
Lung
disease
Tobacco use
Non Tobacco use
No lung
disease
Lung
disease
No lung
disease
Forward in time
Cohort Studies are prospective studies
- At the beginning of the study, subjects either smoke or don’t smoke
- Smokers are put in one group, and non-smokers are put in a second group
- The groups are studied prospectively for any number of years
- The number of subjects in each group that develop the disease are measured
*
Nonexperimental Studies:
no treatment given
Cohort: Subjects don’t have condition but are exposed
Followed over time
Compared with non exposed group
Retrospective Cohort Studies
- In a retrospective cohort study, or historical cohort study, the cohort is identified from past records and followed from the time of those records up to some defined point in the recent past or up to the present time
*
All these factors contribute to reducing bias and increase validity
Prospective
See “Glossary” on Blackboard site
Intervention Trial
- Study conducted on a population in which the independent variable(s) is/are either added or removed and the outcome (dependent variable) is measured subsequently
- Note that the group in this case is a whole population such as a town, subdivision, classroom, or water distribution area
- Individuals within the group are not specifically selected or randomly assigned like those in a clinical trial
*
All these factors contribute to reducing bias and increase validity
Prospective
See “Glossary” on Blackboard site
Nonrandomized Clinical Trial
Nonrandomized Controlled Trial
- May use historical controls
- Subjects receive a new treatment but they are compared with controls (subjects that did not receive the treatment) from another study
- The control and experiment group are not necessarily equivalent
- Can be used to screen a new therapy
*
All these factors contribute to reducing bias and increase validity
Prospective
See “Glossary” on Blackboard site
Randomized Controlled Trial (RCT)
- Trial or study where subjects are randomly assigned to at least two groups
- The control group receives no treatment, the standard treatment, current treatment or placebo
- The experimental group(s) receive the new treatment(s)
- Both groups are followed at the same time
- Strongest level of evidence to demonstrate cause and effect – the treatment caused the effect and the effect did not happen by chance
*
All these factors contribute to reducing bias and increase validity
Prospective
See “Glossary” on Blackboard site
Systematic Review
- Summary of multiple primary research articles that focus on one topic
- Defines and attempts to answer one specific clinical question
- Uses defined criteria in the selection and review of articles
- Reviews RCT and other well-controlled studies
*
Meta Analysis
- Systematic review with the addition of statistical analysis
- Combines data from multiple individual primary research studies into one statistically analyzed study
*
Primary Research
Secondary Research
Individual Research Studies
Reviews of Primary Research
Study 1
Study 2
Study 3
Study 4
Synthesized Results
Statistical Analysis of Synthesized Results
Systematic Review
Meta Analysis
*
Copyright permission granted by SUNY Downstate Medical Center,
Medical Research Library at Brooklyn
Study Designs and Levels of Evidence
Level 1
Level 2
Level 3
Level 4
Level 5
RCT
HIGH
MEDIUM
LOW
*
Lower 2 levels are not ranked: Why not?
As one progresses up the levels,
- the number of studies and the amount of available literature decreases
- at the same time their relevance to answering clinical questions increases
Important in conducting an evidence-based search of the literature
- Knowing which type of research study provides the strongest level of evidence for the question being asked
What if only available research is a case report? What do you tell your patient?
In general:
- The higher you go on the levels of evidence pyramid, the fewer articles there will be on one topic
- The higher you go on the levels of evidence pyramid, the better the evidence is on one topic
*
If the PICO is a therapy or prevention question
Goal: Select treatments that do more good than harm and that are worth the effort and cost
Evidence should consist of:
- Meta analysis (1)
- Systematic reviews (1)
- Single randomized controlled trial (1)
*
If the PICO is a diagnosis question
Goal: to select and interpret diagnostic methods or tests to determine the degree to which a test is useful
Evidence should consist of:
- Meta analysis (1)
- Systematic reviews (1)
- Prospective cohort study (2)
- Single controlled trial that compares new tests with the gold standard (1)
*
If the PICO is an etiology, causation, or harm question
Goal: to identify causes of a disease or condition and to determine relationships between risk factors and disease
Evidence should consist of:
- Meta analysis (1)
- Systematic reviews of RCTs (1)
- Cohort studies (2)
*
If the PICO is a prognosis question
Goal: to estimate the clinical course or progression of a disease or condition over time and anticipate likely complications
Evidence should consist of:
- Meta analysis (1)
- Systematic reviews (1)
- Cohort study (2)
*
Are the other study designs bad?
- Not necessarily
- Depends on your PICO question
- Depends on your patient’s needs
- Aim for high levels of evidence
- Justify why you chose low levels of evidence
*
Literature Reviews
- A literature review is a summary of the published articles relevant to a particular field or topic. It is one person’s evaluation of available information.
- Can be part of a research dissertation/thesis, a research proposal, or a stand-alone publication
- NEVER USE LITERATURE REVIEWS FOR EBDM!!
*
Where is the study design?
- Title
- Abstract Methods section
- Last paragraph of the Introduction
- Methods
- Unlikely to see a “level of evidence” in the article
*
