Chapter 12 Flashcards
An experimental study (intervention study) does what?
Assigns participants to intervention and control groups in order to test whether an intervention causes an intended outcome.
What is the primary distinction between an experimental study and other study designs?
Experimental studies assign participants to receive a particular exposure.
Which three studies are observational designs? (Dont “do” anything)
1) Cross-sectional
2) Case-control
3) Cohort
Which study is the gold standard for assessing causality?
Experimental study
A typical experimental study design in the health sciences is:
A Randomized Controlled Trial (RCT)
What is a randomized controlled trial (RCT)?
1) Some participants are randomly assigned to an active intervention group
2) The remaining participants are assigned to a control group
3) All participants from both groups are followed forward in time to see who has a favorable outcome and who does not.
What is a requirement of an experimental study?
The experiment has to be ethically justifiable.
What do you have to watch out for in experimental studies?
Noncompliance
What is the key statistical measure for experimental studies?
Efficacy
All experimental studies require careful definitions of:
1) The intervention
2) What type of control is appropriate
3) How participants will be assigned to exposure groups
4) What constitutes a favorable outcome for the trial
What is the first step in an experimental study?
Carefully defining the intervention that participants assigned to the active intervention group will receive, and deciding on the person, place, and time (PPT) criteria for the study.
What should the description of the experimental study state?
1) What the intervention will be
2) The eligibility criteria for participants
3) Where and how participants will receive the intervention
4) When, how often, and for what duration participants will receive the intervention 
What are the three types of experimental trials?
1) Superiority trial
2) Noninferiority trial
3) Equivalence trial
What is a superiority trial?
Demonstrates that a new intervention is better than the control. (Must define what “better” is)
What is a non-inferiority trial?
The intervention is not worse than the control.
What is an equivalence trial?
The intervention is equal to the control.
What is the first step for defining the measures of success?
Defining a favorable outcome for an individual.
What is one commonly used type of control group in experimental studies?
Placebo
What is a placebo?
An inactive comparison that is similar to the therapy being tested.
What is the goal of a placebo-controlled study?
It allows the effect of the active therapy to be examined separately from the boost in health status that people may experience simply by participating in a clinical trial or receiving some other intervention.
Do all experimental studies use placebos?
No; if the goal is to test whether a new therapy is better than the current one, we use the current therapy as the control.
Different interventions are compared in various combinations (ex. dosage and duration) within one randomized control trial using what kind of design?
Factorial design.
Why shouldn’t experimental studies include a control group of participants who are randomly assigned to maintain their usual routines?
1) Ethically: discourages the adoption of healthier lifestyles during the course of the study.
2) Hawthorne effect bias: participants change their behavior for the better because they know they are being observed.
Which study uses the same individuals as the control group?
A before-and-after experimental study.
When do we use a before-and-after study?
When there are ethical concerns about the appropriateness of not assigning all participants to potentially life-saving intervention.
What is a before-and-after study
A non-randomized experimental study that measures the same individuals before and after an intervention.
What is the crossover design in a before-and-after study?
Some participants are first assigned to the active intervention and then the control, and others first to the control and then to the active intervention.
Why should we give both groups of the crossover design a break between the two parts of the experiment? What is the break called?
To reduce the carryover effects of the first treatment biasing the apparent results of the second treatment; washout period.
Why are the results of crossover experimental designs not as clear-cut as placebo studies?
Because time alone can lead to significant improvements or declines in health status, especially among the severely ill.
What is blinding or masking?
A type of observer bias where participants in an experimental study (and maybe some research team members) do not know whether a participant is in the active intervention group or the control group.
What is a single-blind study?
A study where participants are unaware of their exposure status.
What is a double-blind study?
A study where neither the participants, nor the people assessing the participants health status know which participants are in the active and control groups.
What kind of bias does blinding minimize?
Information bias that can occur if participants or assessors are able to evaluate outcomes differently based on the results they expect for an exposure. (observer bias if assessors)
When is a blind study only possible?
When all participants are assigned to similar exposures (ex. pill that looks the same).
How can we minimize the likelihood of bias in studies that are not blinded?
By identifying objective outcome measures (such as laboratory tests) rather than subjective outcome measures (such as participants self-reported feelings).
What type of bias does randomization prevent?
Allocation bias: When people with different backgrounds (such as different exposure histories) are not equally distributed across treatment arms.
What type of bias does randomization not remove?
Selection bias: Occurs when people who volunteer to participate in a study are not representative of the source population as a whole
What are the three approaches used to randomly allocate participants to an active intervention group or control group?
1) Simple randomization
2) Block randomization
3) Stratified randomization
What is simple randomization?
Uses a coin toss, a random number generator, or some other uncomplicated procedure to assign each individual to one of the groups.
What is block randomization?
Randomly assigns groups of people to an intervention group and other groups of people to a control group.
What is stratified randomization?
Randomly assigns individuals within certain subgroups to a particular exposure. This ensures that enough members of certain subgroups are being randomized.
What are quasi-experimental designs?
They assign participants to an intervention or control group using a non-random method. (Methods are similar to those of randomized studies).
What are natural experiments?
When the researchers do not have any control over the interventions (ex. Natural disasters).
Are natural experiments true experimental studies?
No, because the interventions cannot be manipulated by a researcher. These questions can be examined with observational study approaches instead.
What ethical issues must be considered before initiating an experimental study?
1) The principle of equipoise
2) The prinicple of distributive justice
3) The principle of respect for persons
4) The principles of beneficence and nonmaleficence
What does the priniciple of equipoise state?
That experimental research should be conducted only when there is genuine uncertainty about which treatment will work better.
What does the principle of distributive justice state?
That the source population must be an appropriate one and that the research must not exploit individuals from populations that are unlikely to have continued access to therapy if it is found to be successful.
What does the principle of respect for persons state?
That all participants volunteer for a study without being unduly influenced by the prospect of being compensated for their participation. Respect also requires that all participants understand what it means to be a research subject, including the possibility of being assigned to a control group instead of the new intervention. (Autonomy)
What do the principles of beneficence and nonmaleficence state?
That researchers balance the likely benefits and risks of the study.
What is an adverse event?
A negative reaction to an intervention or another bad outcome related to a study.
True or false:
Research ethics committee review is required for all experimental studies.
True.
What is the difference between experimental and cohort studies when referring to statistical measures?
Cohort studies: Use these measures to examine the impact of an unassigned exposure on the incidence of disease.
Experimental: use these measures to quantify the impact of an assigned exposure on the likelihood of having a favorable or an unfavorable outcome.
What are the statistical measures that experimental studies use?
1) Rate ratios (RR)
2) Attributable risks (excess risk or risk reduction)
3) Attributable risk percentages
4) Measures of survival
5) Various types of regression models
6) Efficacy
7) Effectiveness
8) Number needed to treat (NNT)
9) Number needed to harm (NNH)
What is efficacy?
The proportion of individuals in the control group who experience an unfavorable outcome who could’ve been expected to have a better outcome if they had been assigned to the active group instead of the control group.
A high efficacy is an indicator that an intervention is ____.
Successful
Efficacy refers to results under:
Ideal circumstances (like full compliance. Lab based)
What is effectiveness?
It is calculated with the same equation as efficacy, but refers to results obtained under less than ideal circumstances. (Less compliance. “Real world” setting)
What is the number needed to treat (NNT)?
The expected number of people who would have to receive a treatment to prevent an unfavorable outcome in one person (or to achieve a favorable outcome in one person).
What does a number needed to treat (NNT) of 5 mean?
That five people have to receive a treatment in order for one to prevent an unfavorable outcome.
What is the number needed to harm (NNH)?
The number of people who would need to receive a particular treatment in order to expect that one of them would have a particular adverse outcome.
A safe intervention is expected to have a:
Large NNH
Which two statistical measures are often used for cost-effectiveness analysis?
NNT and NNH
What is efficiency?
An evaluation of the cost effectiveness of an intervention that is based on both effectiveness and resource considerations.
What is a treatment-received approach?
It limits analysis to the participants who were fully compliant with their assigned intervention
What is a treatment-assigned approach (intention-to-treat approach)?
Includes all participants in analysis even if they were not fully compliant with their assigned intervention.
Which analysis allows for the calculation of efficacy and which allows for the calculation of effectiveness?
Treatment-received; Treatment-assigned (intention-to-treat)
What should be included in reports of findings for any experimental study?
The flow of participants through the study, from the recruitment and enrollment stages through the analysis stage.
The research protocol for any experimental study should include specific plans for:
1) Promoting compliance
2) Minimizing dropouts
What is the sensitivity?
The proportion of people who actually have a disease (according to the reference standard) who test positive using the new test.
What is the specificity?
The proportion of people who do not have the disease who test negative with the new test.
What is the positive predictive value (PPV)?
The proportion of those who test positive with the new test actually have the disease (according to the reference standard).
What is the negative predictive value (NPV)?
The proportion of those who test negative who actually do not have the disease.
A good screening or diagnostic test will have
High values for their measures.
If the cut off values change which measures change?
All 4 diagnostic and screening measures.
What do receiver operating characteristic curves or ROC curves plot?
X-axis: (1-specificity)
Y-axis: (sensitivity)
What do ROC curves and the area under the curve (AUC) examine?
The accuracy of a diagnostic test.
What are the three other measures commonly used for screening tests?
1) Diagnostic accuracy
2) Positive likelihood ratio test (LR+)
3) Negative likelihood ratio test (LR-)
What is the diagnostic accuracy?
The percentage of the participants who are either true positives or true negatives. {Percentage for which both the reference test and the new test yield the same result}
An ideal diagnostic accuracy test will have:
100% diagnostic accuracy.
What is the positive likelihood ratio test?
Examines whether a new test is good at predicting the presence of disease.
What is the formula for the positive likelihood ratio test?
LR+= (sensitivity)/(1-specificity)
{The probability that an individual with the disease has a positive test divided by the probability that an individual without the disease has a positive test}
An LR+ larger than ____ indicates a good test.
LR+ > 10
What is the negative likelihood ratio test?
Examines whether a new test is good at predicting the absence of disease.
What is the formula for the negative likelihood ratio test?
LR-=(1-sensitivity)/(specificity)
{The probability that an individual with the disease has a negative test divided by the probability that an individual without the disease has a negative test}
An LR- smaller than ____ indicates a good test.
LR- < 0.1
