Term
| Describe cross-sectional study. |
|
Definition
Subjects sampled at fixed point in time and investigate association between presence or absence of Rick factors and disease.
Strength- inexpensive, population based
Weakness- recall bias |
|
|
Term
|
Definition
Tracks patients with known exposure and similar treatment or examine records of exposure and outcome.
Strengths- investigate uncommon diseases, inexpensive, hypothesis-generating
Weakness- no control group, no statistical validity, may have selection bias |
|
|
Term
| Describe case control study. |
|
Definition
Compare patients with condition to patients without condition.
Strengths- inexpensive, quick study, investigate infrequent events, useful in generating odds ratio
Weaknesses- Do not give absolute risk, may have confounded, does not show cause and effect |
|
|
Term
|
Definition
Longitudinal study, cohort determined prior to appearance of condition
Strengths- data over regular intervals, reduced recall error, generate relative risk
Weaknesses- expensive, lengthy follow up |
|
|
Term
| Describe randomized control trial. |
|
Definition
Randomized allocation of different interventions
Strengths- consistent selections, randomizations remove most bias
Weaknesses- expensive, loss of follow up, patients may not be compliant |
|
|
Term
Give examples of parametric tests (dependent variables symmetric or variance close)?
Non parametric tests (skewed or not close)? |
|
Definition
Student t-test, ANOVA, paired t-test, least square correlation
Wilcoxon rank-sum test, kruskal-wallis test, wilcoxon signed-rank test, spearman rank-order correlation |
|
|
Term
|
Definition
Null hypothesis falsely rejected, p value is chance of accepting difference when there is none
If p<0.05 then less 5% chance of type I error |
|
|
Term
|
Definition
Fail to reject null hypothesis when it is not true
Common cause is too small sample size |
|
|
Term
| Describe statistical power. |
|
Definition
% chance difference will be detected if one exists
1-% chance of type II error
Ie. if type II error is 5% then power is 95% |
|
|
Term
| Describe normal distribution. |
|
Definition
| Bell curve- 68% within 1sd, 95% within 2sd |
|
|
Term
Describe odds ratio.
Risk ratio.
Prevalence.
Incidence. |
|
Definition
Probability of occurrence to that of nonoccurrence, true tests/false tests
Ratio of risk of those exposed to those not exposed
% of those with condition (presence of condition)
Number of new disease cases over time interval |
|
|
Term
Describe sensitivity.
Specificity.
PPV.
NPV. |
|
Definition
Fraction of all true cases the test detects, true positives or number with disease, associated with false negative rate, rule out, true positives/all positives
Fraction of all negative cases the test detects, true negatives, associated with low false positives, rule in, true negatives/all negatives
Probability that positive test indicates disease, true positives/true positives + false positives
Probability that negative test indicates no disease, true negatives/true negatives + false negatives |
|
|
Term
| Describe levels of evidence. |
|
Definition
Level I at least one RCT
Level II-1 controlled trials without randomization
Level II-2 cohort or case-control studies from different locations
Level I-3 multiple time series or uncontrolled trials
Level III opinions |
|
|
Term
| Describe levels of recommendations. |
|
Definition
A good scientific evidence
B fair evidence
C fair evidence but balance of risk and benefit too close to generalize
D fair evidence risks outweigh benefits
I no evidence |
|
|
Term
| Describe elements of consent. |
|
Definition
| Competence, disclosure, understanding, voluntariness, consent |
|
|
Term
| What requires mandatory reporting? |
|
Definition
| OTC products, dietary supplements, drug/biological/human cell tissue and cellular/tissue based product manufacturers/distributors/packers, human cell or tissue products |
|
|
