Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
70 Cards in this Set
- Front
- Back
Validity
|
does it measure what we think it does
|
|
Reliability
|
repeatability, will it give same result over and over (at different times and with different examiners?
|
|
What is the best test for reliability?
|
Kappa
|
|
What does a high kappa show?
|
Better reliability, don’t use a test with kappa less than .60
|
|
How do you evaluated diagnostic procedures?
|
Validity and reliability
|
|
How do you assess validity of a diagnostic?
|
Blind comparison with a gold standard
|
|
How do you do a blind comparison with a gold standard?
|
ID groups of sick and healthy, administer a new test in a blinded manner, calculate percent of correctly identified subjects with regard to gold standard
|
|
Sensitivity
|
the probability of a positive test result when the disease is present
|
|
Specificity
|
probability of a negative test result when the disease is absent
|
|
PPV
|
probability that the patient has the disease when the test is positive
|
|
NPV
|
probability that the patient does not have the disease when the test is negative
|
|
Effect of disease prevalence
|
as the prevalence of a disease decreases the PPV of the diagnostic test gets worse (an increase in false positives)
|
|
Criteria for acceptance of a new diagnostic
|
has safety, validity, and reliability been shown? Does it add anything to the diagnostic test which precedes it? Advantages in cost/speed/comfort? Will it cause you to change your treatment plan?
|
|
Prospective research study
|
designed to answer specific questions about biomedical or behavioral interventions (drugs, treatments, devices, or new ways of using known drugs, treatments, or devices
|
|
Clinical trials
|
used to determine whether new biomedical behavioral interventions are safe, efficacious, and effective
|
|
Efficacy
|
does the treatment work under ideal conditions?
|
|
What is the focus on for efficacy?
|
Internal validity
|
|
Internal validity
|
the integrity of the experiment itself, ability to draw a causal link between your treatment and the dependent variable of interest
|
|
Effectiveness
|
does the tx work under real-life conditions? |
|
What is the focus on for effectiveness?
|
External validity
|
|
External validity
|
ability to generate your study findings to the population at large
|
|
Things bias can affect
|
trend of results, internal and external validity
|
|
Confounding variables
|
some aspect of a subject that is associated with both the outcome of interest and with the intervention of interest
|
|
What do confounding variables produce?
|
Alternative, competing explanations for an effect
|
|
what can you do to reduce confounding variables?
|
Measure baseline characteristics of subjects at initial recruitment (prognostic factors—smoking, medications, hx of periodontal disease)
|
|
examples of confounds
|
noncompliance, tx received outside the study, pt on placebo receives active tx, incorrect timing of follow up, incorrect dx, biased assessment of endpoint, biased admin of care
|
|
how can you control confounders?
|
Holding confounding variable constant, rely on randomization, matching (similar intervention groups w respect to important prognostic factors), stratification (division of participants into subgroups to categorical prognostic factors), post-study subgroup analysis (no fishing), statistical control (analysis of covariance)
|
|
what is the gold standard of study designs?
|
RCT
|
|
how are RCTs designed?
|
To minimize bias through randomization and blinding—internal validity |
|
what can RCTs provide?
|
Sound evidence of causation
|
|
how does crossover design work?
|
Each patient is his or her own control, receive tx, washout, next tx
|
|
how does split mouth work?
|
Each pt own control, diff tx on each side of mouth or quad
|
|
random sampling
|
every member of the population has an equal and independent chance of being selected
|
|
convenience sampling
|
take every patient from the accessible population who meets the selection criteria over the specified time period
|
|
stratified sampling
|
if the outcome being measured is rare in populations as a whole, recruit at random or consecutively from populations at high risk of the condition in question
|
|
power of a statistical test
|
probability that the study sample will yield statistically significant results when there truly is a treatment effect (aka clinically important difference) in the target population
|
|
statistical significance
|
the observed difference is greater than zero and unlikely to have occurred by chance alone
|
|
clinical significance
|
the smallest difference that would make a meaningful difference in patients lives or clinicians practice
|
|
alpha level
|
significance level, or probability of rejecting the null hypothesis when the null hypothesis is true
|
|
simple randomization
|
all subjects have an equal and independent chance of being allocated (must truly be a chance process)
|
|
what do control groups allow for?
|
Isolation of patient outcomes caused by the test treatment from outcomes caused by other factors
|
|
analysis by intention to treat
|
all subjects are analyzed in the groups to which they were assigned, regardless of what happened in the study, tests the effects of assigning patients to treatments (external validity)
|
|
analysis by treatment received
|
tests the effects of administering treatments
|
|
registration of RCT
|
establishes baseline for judging quality of trial, must be registered prior to consideration for publication, register before you recruit
|
|
CONSORT
|
consolidation standards of reporting trials
|
|
Phase 0
|
pharmacodynamics/kinetics, first human trials 10-15 subjects
|
|
Phase 1
|
20-80 subjects, determine tolerance, metabolism ,kinetics,bioavailability, toxicity and major SEs, establish a safe dose range, relatively high risk
|
|
Phase 2
|
establish the testing protocol, 100-300 subjects, gather info on effectiveness on actual clinical endpoints in the targeted indication, ID MC sort term SEs and risks, determine dose ranges for phase 3
|
|
Phase 3
|
1000-3000 subjects, adequate well controlled study (RCT), demonstrate efficacy and safety, overall risk/benefit, provide info for labeling
|
|
Phase 4
|
post marketing surveillance, ID less common and long term SEs, compare to other marketed products, study new patient populations or new indications
|
|
Power analysis
|
preferred method to determine number of subjects in your study
|
|
Type of study with the poorest internal validity
|
case study
|
|
Type of study with the best external validity
|
meta analysis
|
|
Narrative review
|
broad topic, few inclusion criteria, susceptible to bias, non-reproducible
|
|
Systematic review
|
narrow topic, exhaustive lit search, strict inclusion/exclusion
|
|
Dimensions of quality in an SR
|
clearly stated question, search strategy, things in duplicate, inc.excl, quality of included studies, results
|
|
Why perform a meta analysis?
|
Increases statistical power, improve precision, settle controversies from conflicting studies or generate new hypotheses
|
|
What do M-A’s do?
|
Derive meaningful conclusions from data and help prevent errors in interpretation
|
|
Heterogeneity
|
any kind of variability among studies (clinical—participants, interventions; methodologic—trial design, quality; statistical)
|
|
Measurements of heterogeneity
|
Q test (chi squared) or I^2 (for quantifying inconsistency across included studies—0-40% might not be impt, 30-60 moderate, 50-90 substantial, 75-100 considerable)
|
|
Standardized mean difference
|
used when trials assess the same outcome but measure in a variety of ways, including different scales
|
|
Risk ratio
|
risk/prob/chance of the occurrence of an event in tx relative to control
|
|
Odds ratio
|
odds of an even occurring to it not occurring for tx relative to control
|
|
interpretation of a chi square test |
a large chi squared statistic relative to its degrees of freedom provides evidence of heterogeneity |
|
what is an I^2 used for |
quantifying inconsistency among included studies |
|
NNT |
number of pts needed to treat to prevent one bad outcome |
|
how does CI relate to sample size |
inverse sqrt ratio, therefore smaller samples generate wider intervals, if you want to cut margin of error in half then quadruple the SS |
|
internal validity associated with effectiveness or efficacy |
efficacy |
|
internal validity related to the experimental conclusions |
the ability to draw a causal link between your treatment and the dependent variable of interest |
|
clinical trials are used to... |
determine whether new biomedical or behavioral interventions are safe, efficacious, and effective |