Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
WEEK 4:
A snapshot in tine - cross sectional studies:
| Question | Answer |
|---|---|
| cross sectional study | gather information on a sample of people in order to estimate characteristics of larger population (individual level data in a single point in time) |
| examples of a cross sectional study (3) | questionnaire, telephone interview and mail shots |
| strengths of cross sectional studies (2) | collects lots of data quickly and relatively cheap and simple to do with no follow up |
| limitations of cross section studies (3) | generalisability (can results of study in sample be applied to larger population), no useful for rare exposures or diseases (would have to survey many people), and cant reliably assess causation (only snapshot in time) |
| main measures (Ratio) (2) | odds ratio (OR) or risk ratio (RR) |
| if ratio = 1 | no difference in odds/ risk between groups (null hypothesis) |
| if ratio <1 | exposed group is less likely to have outcome |
| if ratio >1 | exposed group is more likely to have outcome |
| relative risk (risk ratio) | intuitive and easy to calculate and understand (but cant be calculated for case control studies) |
| odds ratio | harder to understand but CAN be calculated for case control studies |
| when does odds ratio = risk ratio (2) | if disease/ event is rare or if the OR/RR is close to 1 |
| what does a ratio of 2 mean | twice as likely to have disease in exposed group |
| what does a ratio of 0.3 mean | 70% less likely to have disease in exposed group |
| what does a ratio of 1.2 mean | 20% more likely to have disease in exposed group |
| attributable risk | risk in exposed group - risk in unexposed group |
| categorical data | ordinal (eg cancer stage I-IV or likert scale (strongly agree, agree, disagree, strongly disagree etc) or nominal (choice of treatment eg medication/surgery/exercise) |
| continuous data | any value within an interval (eg weight, temperature, BMI, age) |
| binary data | yes/no or dead/alive etc (2 straightforward answers) |
| mean difference formula | exposed - unexposed |
| if mean difference = 0 | no difference in effect between groups (eg value expected if null hypothesis is true) |
| if mean difference <0 | outcome is lower in exposed group |
| if mean difference >0 | outcome if higher in exposed group |
| statistical inference | going beyond the study to reach a conclusion about a larger population (sample small population to apply it to a bigger population) |
| measure of effect for statistical inference | directly uses sample estimate to apply to population BUT how confident are we about reliability of this estimate eg sampling error, sample size |
| confidence interval for statistical inference includes | likely range that true measure of effect lies within, range of true values consistent with observed data, sample measure of effect, allows to judge statistical significance (alternative to p value) |
| what happens to confidence interval when sample becomes bigger | smaller (because you have more samples to narrow down the true value- more accurate result) |
| confidence interval takes into account | random variation (bias) |
| importance of confidence interval | gives additional information about how confident we are that intervention actually has a beneficial effect and likely size of that effect (more informative than p value alone) |
| p<0.05 | there is a less than 5% chance we would see results this extreme if H0 is true and there is no real association |
| 95% confidence interval | range of values that we are 95% confident the true effect lies within |
| if 95% confidence interval includes null hypothesis then p is | >0.05 |
| if 95% confidence interval does not include null hypothesis then p is | <0.05 |
| if H0 (RR=1) lies within 95% confidence interval then H0 value is | consistent with observed data and p>0.05 so we can not reject H0 (null hypothesis) |
| if H0 (RR=1) lies outside 95% confidence interval then H0 value is | NOT consistent with observed data and p<0.05 so we CAN reject the H0 (null hypothesis) |
Created by:
22tango
Popular Medical sets