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
USMLE2 Epi
Epidemiology 1
| Question | Answer |
|---|---|
| What is prevalence? | number of existing cases in the population at a moment in time |
| What is incidence? | number of new cases in the disease-free population that develop over a period of time |
| Formula for prevalence? | Prevalence = incidence x average duration of disease; also = (total cases in population at a given time)/(total population) |
| incidence is measured with what type of study | cohort study |
| prevalence is measured with what type of study? | prevalence (cross-sectional) study |
| What is a cross-sectional study? | study in which people in a population are examined for the presence of a disease of interest at a given point in time |
| Advantages of cross-sectional studies (3). Disadvantages of cross-sectional studies (2). | Adv: 1. efficient to look at population (cases & non-cases at same time), 2. basis for dx testing, 3. plan health services to offer. Disadv: 1. can't determine causality (only one point in time), 2. risk & incidence of dz can't be directly measured |
| What statistical test to estimate causal relationships? | chi squared |
| What statistical test to estimate relative risk or exposure? | odds ratio |
| Formula for Incidence? | Incidence = (NEW cases in population over a given time period)/(total population at risk during that time); Note: when calculating incidence don't forget that people peviously positive for a disease are no longer considered at risk. |
| When is prevalence > incidence? | chronic diseases (e.g., diabetes) |
| When is prevalence = incidence? | acute diseases (e.g., common cold) |
| Ratio for false negative | 1 - sensitivity |
| Ratio for false positive | 1 - specificity |
| Sensitivity - what is it and what is the formula? | a=dz with + test; b=false +; c=false neg; d=real neg result. Sensitivity is the number of TRUE POSITIVES divided by the number of all people with the disease, the probability of a POSITIVE TEST given that a person has the disease. a/(a+c) |
| Specificity - what is it and what is the formula? | a=dz with + test; b=false +; c=false neg; d=real neg result. Specificity is the number of TRUE NEGATIVES divided by the number of all people without the disease, probability of a NEGATIVE TEST given that a person does not have dz. |
| PPV - what is it and what is the formula? | a=dz with + test; b=false +; c=false neg; d=real neg result. Prob that a pt with a pos result actually has the disease. PPV = a/(a+b) |
| NPV - what is it and what is the formula? | a=dz with + test; b=false +; c=false neg; d=real neg result. Prob that a pt with a neg result does NOT have the disease. NPV = d/(c+d) |
| What will give a higher positive predictive value? | greater specificity or higher prevalence will give a better PPV |
| What will give a higher negative predictive value? | greater sesitivity or lower prevalence will give a better NPV |
| Number of true positives divided by the number of people who tested positive for the disease? | Positive Predictive Value (PPV) |
| The probability of having a condition given a positive test? | Positive Predictive Value (PPV) |
| The number of true negatives divided by the number of people who tested negative for the disease? | Negative Predictive Value (NPV) |
| The probability of not having the condition given a negative test? | Negative Predictive Value (NPV) |
| Unlike sensitivity and specificity, predictive values are dependent on the WHAT of the disease. | Unlike sensitivity and specificity, predictive values are dependent on the PREVALENCE of the disease. |
| What is a likelihood ratio? | how much more or less likely a given test result is in diseased vs. non-diseased people |
| what is a pretest probability | the prevalence of a disease |
| what do you use a likelihood ratio for? What is a +LR? What is a -LR? | provides a direct estimate of how much a test result will change odds of having dz. LR for a + result tells you how much the odds of dz increase when a test is +. The LR for a - result (LR-) tells you how much the odds of the dz when a test is negative. |
| formula for + likelihood ratio | real positive/false positive = sensitivity/(1-specificity) |
| formula for - likelihood ratio | false negative/real negative = 1-sensitivity/(specificity) |
| What does it mean if a test has an LR of 1? 10? 0.1? | LR 1 means the pretest prob that the pt has the disease does not change with the test; LR 10 means that disease is 45% more likely; LR of 0.1 means that disease is 45% less likely |
| What is a post-test odds? What is the formula? | the chances that your patient has a disease, given the test result. Formula is Pre-test odds (individualized) x likelihood ratio. |
| Odds Ratio (OR)? | Odds of having disease in exposed group divided by odds of not having disease in exposed group. |
| For Odds Ratio, odds are calculated xxx as the number with disease divided by the number without disease. | For Odds Ratio, odds are calculated WITHIN A GROUP as the number with disease divided by the number without disease. |
| In what situation does Odds Ratio (OR) approximate Relative Risk? | if prevalence of disease is not too high. |
| Odds Ratio is used for xxx studies. | Odds Ratio is used for CASE-CONTROL studies. |
| What is the absolute risk? | incidence of disease |
| Formula for Odds Ratio? | a=+dz+exp, b=-dz+exp, c=+dz-exp, d=-dz-exp; OR = (a*d)/(b*c) |
| Formula for Relative Risk? | a=+dz+exp, b=-dz+exp, c=+dz-exp, d=-dz-exp; RR is incidence in the exposed over incidence in the unexposed. RR = a/(a+b) divided by c/(c+d) |
| Formula for Attributable Risk? | AR = a/(a+b) minus c/(c+d). Additional incidence of disease that is due to a risky exposure, on top of the background incidence |
| Relative Risk (RR)? | Disease risk in exposed group divided by disease risk in unexposed group. This is how much more likely an exposed person is to get dz compared to unexposed person. Indicates the strength of the association between the exposure and the disease. |
| Risk is calculated xxx as the number with disease divided by the total number of people in the group. | Risk is calculated WITHIN A GROUP as the number with disease divided by the total number of people in the group. |
| Relative Risk (RR) is used for xxx studies. | Relative Risk (RR) is used for COHORT studies. |
| To commit a Type I error (alpha) is to state what? | FALSE POS. There IS an effect or difference when none exists (to mistakenly accept the experimental hypothesis and reject the null hypothesis). |
| p is judged against xxx, a preset level of significance (usually < 0.05). | p is judged against alpha, a preset level of significance (usually < 0.05). |
| p = ? | p = probability of making a type I error. |
| If p < 0.05, then there is less than a 5% chance that xxx. | If p < 0.05, then there is less than a 5% chance that THE DATA WILL SHOW SOMETHING THAT IS NOT REALLY THERE. |
| Layman's way of describing alpha? | alpha = you "saw" a difference that did NOT exist--for example, convicting an innocent man. |
| In a four quadrant box, power lies in what region? | Power is at the intersection of column H1 (reality) and row H1 (study results) |
| In a four quadrant box, alpha lies in what region? | Alpha is at the intersection of column H0 (reality) and row H1 (study results) |
| In a four quadrant box, beta lies in what region? | Beta is at the intersection of column H1 (reality) and row H0 (study results) |
| To commit a Type II error (beta) is to state what? | FALSE NEG. There is NOT an effect or difference when one exists (to fail to reject the null hypothesis, when, infact H0 is false). |
| Beta is the probability of making a type xxx error. | Beta is the probability of making a type II error. |
| Layman's way of describing beta? | Beta = you did not "see" a difference that does exist--for example, setting a guilty man free. |
| Qualitative definition of Power? What is its formula? | Power is the probability that a study will find a statistically significant conclusion when there really is one there. Power = 1 - type II error. |
| Power depends upon what (3 items)? | 1. Total number of end points experienced by population. 2. Difference in COMPLIANCE b/w treatment groups (differences in the mean values b/w groups). 3. Size of expected effect. |
| If you xxx sample size, you increase Power. | If you INCREASE sample size, you increase Power. There is Power in numbers. |
| Formula for SEM? | SEM = SD/(square root of sample size) |
| SEM xxx SD? | SEM < SD? |
| SEM xxx as sample size increases? | SEM DECREASES as sample size increases? |
| For a Normal (Gaussian) distributional curve, SD of 1 = x%? | SD 1 = 68% |
| For a Normal (Gaussian) distributional curve, SD of 2 = x%? | SD 2 = 95% |
| For a Normal (Gaussian) distributional curve, SD of 3 = x%? | SD 3 = 99.7% |
| CI = range from xxx to xxx? | CI = range from [mean - Z(SEM)] to [mean + Z(SEM)] |
| The 95% CI corresponds to what p value? | p = 0.05 |
| For the 95% CI, Z = xxx. | For the 95% CI, Z = 1.96. |
| If the 95% CI for a xxx between 2 variables includes 0, then there is no significant difference and H0 is NOT rejected. | If the 95% CI for a MEAN DIFFERENCE between 2 variables includes 0, then there is no significant difference and H0 is NOT rejected. |
| If the 95% CI for xxx or xxx includes 1, then H0 is NOT rejected. | If the 95% CI for ODDS RATIO or RELATIVE RISK includes 1, then H0 is NOT rejected. |
| Chi squared checks what? | difference b/w 2 or more percentages or proportions of categorical outcomes (NOT mean values). |
| Chi squared = | compare percentages (%) or proportions |
| r squared = | Coefficient of determination |
| Mnemonic for reportable diseases IN ALL STATES? | B.A. S.S.S.M.M.A.R.T. Chicken or you're Gone: Hep A, Hep B, Salmonella, Shigella, Syphilis, Measles Mumps, AIDS Rubella, TB, Chicken, Gonorrhea |
| Which disease can vary by state for reporting? | HIV |
| Medicare Part A = | hospital |
| Medicare Part B = | doctor bills |
| what is a cohort study? | longitudinal study/prospective study. Start with a group of people who DON'T have the outcome of interest, but who could potentially develop the outcome of interest. (Group 1 is exposed to some variable while Group 2 is the control.) |
| Advantages of cohort study (3) | 1. can establish incidence/absolute risk, 2. can assess the relationship of exposure to many diseases, 3. no bias from a known outcome |
| Disadvantages of cohort study (3) | 1. expensive, 2. only study the relationship of disease to the exposures recorded at the beginning of the study, 3. requires many subjects --> inefficient and can't study rare diseases |
| What is a case-controlled study | retrospective study. two groups (with disease, without disease) then looks BACK in time to measure the comparative frequency of exposure to a possible risk factor |
| Advantages of case-controlled study (3) | 1. small groups --> less expensive, 2. can study rare diseases, 3. can look at multiple risk factors |
| Disadvantages of case-controlled study (2) | 1. numbers determined artificially since groups were chosen based on outcome (can't calc prevalence, incidence or relative risk, 2. retrospective data may be inaccurate |
| Formula for odds | probability of event / (1-prob of event) |
| formula for probability | odds / (1+ odds) |
| What kind of bias: different groups are assessed/measured using different tools | measurement bias |
| What kind of bias: found in retrospective studies | recall bias from people remembering incorrectly |
| What kind of bias: disease is identified earlier so followed longer - pt not actually living longer | lead-time bias |
| What kind of bias: screening tests detect disproportionate number of progressive diseases but miss rapidly progressing ones | length bias |
| what is the confidence interval? | there is a 95% chance that the interval contains the true effect seize, which is likely to be closest to the point estimate near the center of the interval |
| What does it mean if a CI includes the value corresponding to a relative risk of 1? | the results are not statiscally significant |
| If the CI is wide, then the power is… | low |
| Quality improvement tool that is used to do root cause analysis | Cause and effect diagram – Fishbone or Ishakawa diagram |
| A chart that focuses improvement initiatives on the most common root cause of the problem | a Pareto chart |
| Chart that visually displays flow through the system. Diagram can help highlight inefficiencies or redundancies | Spaghetti diagram in which flows are drawn as lines on a map, I need to follow a medication through a hospital from order generation to administration of the medication. |
Created by:
christinapham
Popular USMLE sets