UWorld Epidemiology

A placebo-controlled clinical trial is conducted to assess whether a new antihypertensive drug is more effective than standard therapy. A total of 5000 patients with essential hypertension are enrolled and randomly assigned to one of two groups: 2500 patients receive the new drug and 2500 patients receive placebo. If the alpha is set at 0.01 instead of 0.05, which of the following is the most likely result? (A) Significant findings can be reported with greater confidence (B) The study will have more power (C) There is a decreased likelihood of a Type II error (D) There is an increased likelihood of statistically significant findings (E) There is an increased likelihood of a Type I error

(A) Significant findings can be reported with greater confidence P value is the probability that findings are due to chance alone A statistically significant difference is one that cannot be accounted for by chance alone P values are considered statistically significant when they are less than a set alpha value/significance level Alpha being set to a lower value, means that for statistical significance, P value must be less than a smaller number and less chance that outcome is by chance alone E is wrong because decreasing alpha decreases likelihood of type 1 error

When compared to the 99% confidence interval, the 95% confidence interval is (A) less precise and less accurate (B) more precise but less accurate (C) more precise and more accurate (D) less precise and more accurate

(D) less precise and more accurate CL↑ Width↑ Accuracy↑, but Precision↓ The wider the CI, the less precise and the more accurate the estimate of the mean. When compared to a 99% confidence interval, a 95% confidence interval will be more precise (smaller SE and width of the confidence interval) but less accurate (the sample is less likely to be representative)

Effect Modification

* * * * *

Status of Change Model

* * * * *

Confounding Bias

* * * * * *

Lead time bias

* * * * * *

What are three ways to increase power/ decrease B

* increase sample size * increase expected effect size * increase precision

Type 1 Error

**Stating that there is an effect when none exists* *Null hypothesis is incorrectly rejected in favor of alternative hypothesis **α is the probability of making a type I error* * *

Type 2 Error

**Stating that there is not effect when one exists* *Null hypothesis is not rejected when it is in fact false **b is the probability of making a type 2 error* * *

Types of Studies

*Case series -Report on observations of patients with known exposure or disease -Does not have comparison group so cannot perform hypothesis testing *Case control *Observational (by definition) *Retrospective (almost always)* *Compares the odds of being exposed between patients with disease and patients without the disease *e.g. Patients with cirrhosis are more likely to have been exposed to heavy alcohol use *Odds Ratio (OR) is measure of disease association OR = (a/b) / (c/d) = ad/bc *Cohort *Observational (by definition) *Can be either prospective or retrospective* *Compares a group with a given risk factor to a group without the risk factor to see if there is an increased likelihood of developing disease *e.g. Patients who drink alcohol are more likely to get cirrhosis *Relative Risk (RR) is measure of disease association *asks, "How much more likely are you to get cirrhosis if you drink alcohol?" *RR = [a/(a+b)] / [c/(c+d)] *Cross sectional -Observational -Determines disease status and exposure/risk factor status at the *same point in time* -can show association but not causality -measures disease *prevalence* *Cross over -Participants alternate receiving intervention and placebo -Participants act as own controls -improves power and precision of study -all patients receive intervention *Twin concordance study -Compares monozygotic and dizygotic twins to measure disease heritability Adoption study -Compares siblings raised by biologic and adoptive parents to measure *influence of environmental vs inherited* factors *Meta analysis -Combines data from multiple studies -Better precision than individual studies -Improves generalizability of study findings -Considered to be the highest level of clinical evidence -Limited by: -quality of individual studies -bias in study selection

Chi-square test

*Checks differences between 2 or more percentages or proportions of categorical outcomes *Similar to t-test, but it is used for categorical measures * * *

Analysis of Variance

*Checks differences between means of 3 or more group * * * * * *

T-test

*Compares the means of 2 groups *Requires normally distributed, continuous measurements * *Null hypothesis assumes that there is no difference between the two groups *p<0.05 rejects the null hypothesis = 2 means are statistically different * *

Absolute risk reduction (ARR)

*Event rate in control - event rate in treatment *The difference in risk attributable to the intervention *if 8% of people who receive a placebo vaccine develop the flu vs 2% of people who receive a flu vaccine, then ARR = 8% − 2% = 6%

Relative Risk

*Measures the effect of exposure on the chances of developing a disease *Typically used in cohort studies *Risk of developing disease in the exposed group DIVIDED by risk in the unexposed group * *If prevalence is low, OR ≈ RR. *

Null Hypothesis vs Alternative Hypothesis

*Null (H0)/ Hypothesis of no difference *Ex: no association between the disease and the risk factor *Alternative (H1)/ Hypothesis of relationship *Ex: there is some association between the disease and the risk factor A correct result is rejecting the null in favor of H1 when a difference exists or not rejecting the null when no difference exists

Odd's Ratio

*Odds that the group with the disease (cases) was exposed to a risk factor divided by the odds that the group without the disease (controls) was exposed **OR = A/C / B/D or AD/BC* *"Probability of having a risk factor if one has a disease" *Typically used in case control studies * * * *

Compare the relationship of mean, median, and mode in the following distributions *Positively skewed *Negatively skewed *Normal distribution

*Positively skewed: Mean > median > mode (due to large magnitude values) *Negatively skewed: Mean < median < mode (due to small magnitude values) *Normal distribution: Mean = median = mode

Define statistical power

*Probability of rejecting the null hypothesis when it is false **Or in other words, probability of correctly stating that there is an effect when one exists* *The power of an experiment is the probability that it can detect a treatment effect, if it is present. *(1 - β) * * *

Number needed to harm (NNH)

*Represents the number of people who must be exposed to a risk factor before 1 additional adverse event occurs *NNH = 1/absolute risk increase (AR: attributable risk) *AR = adverse event rate in control - adverse event rate in treatment group

Number needed to treat (NNT)

*Represents the number of people who must be treated before 1 person benefits *NNT = 1/absolute risk reduction (ARR) *ARR = the difference in risk attributable to the intervention as compared to the control

Attributable Risk

*The DIFFERENCE in risk between exposed and unexposed *Or, the proportion of disease occurrences that that are attributable to the exposure *If risk of lung cancer in smokers is 21% and risk in nonsmokers is 1%, then 20% of the lung cancer risk in smokers is attributable to smoking *AR = a/(a + b) -c/(c + d

Relative Risk Reduction

*The PROPORTION of risk reduction attributable to the intervention as compared to a control *RRR = 1 - RR *If 2% of patients who receive flu shot develop the flu, while 8% of unvaccinated patients develop the flu, then RR = 2/8 = 0.25, and RRR = 0.75

Likelihood Ratio

*The likelihood that a test will change the pre-test probability that the condition exists * *Higher ratio = better test *post test = pre test X likelihood ratio * * * * *

Confidence Interval

*The range of values within which the true mean of a population is estimated to lie *If the 95% CI for a mean difference between 2 variables includes 0, then there is no signifcant difference and H0 is not rejected. *If the 95% CI for odds ratio or relative risk includes 1, H0 is not rejected. *If the CIs between 2 groups do not overlap, a statistically signifcant difference exists. If the CIs between 2 groups overlap ,then usually no signifcant difference exists * * *

How does sample size affect a Type 1 vs Type 2 error

*Type 1 error is not affected by sample size *Type 2 error decreases with increase in sample size (beta gets smaller)

What effect does increased precision have on the following parameters? *Confidence interval *Standard deviation *Statistical power

*Width of CI goes down *SD goes down *Power goes up

Pearson Correlation coefficient (r)

*r is always between −1 and +1 *Closer r is to 1, stronger the correlation between the two variables *Positive r means positive correlation (as one variable inc, the other inc *Negative r means negative correlation (as one variable inc, the other dec) * *

When looking at confidence intervals for relative risk, relative benefit, etc, remember that if the CI 95% crosses ________________, there is no difference between the groups.

1

A study on cholesterol levels of a town in rural Idaho is performed, of which there are 1000 participants. It is determined that in this population, the mean LDL is 200 mg/dL, with a standard deviation of 50 mg/dL. If the population has a normal distribution, how many people have a cholesterol less than 300 mg/dL? Topic Review Topic 1. 680 2. 925 3. 950 4. 975 5. 997

3. 975

You would like to conduct a study investigating potential risk factors that predispose patients to develop cirrhosis. Using a registry of admitted patients over the last 10 years at your local hospital, you isolate all patients who have been diagnosed with cirrhosis. Subsequently, you contact this group of patients, asking them to complete a survey assessing their prior exposure to alcohol use, intravenous drug abuse, blood transfusions, personal history of cancer, and other medical comorbidities. An identical survey is given to an equal number of patients in the registry who do not carry a prior diagnosis of cirrhosis. Which of the following best describes the type of study you are attempting to conduct? Topic Review Topic 1. Cohort study 2. Meta-analysis 3. Case-control study 4. Cross-sectional study 5. Randomized controlled trial

3. Case-control study The study described is a retrospective case-control study. In this study you start with a disease (the sick patients) and look for prior exposures to certain risk factors and calculate an odds ratio to analyze the data.

In a normal distribution, +/- 1 SD = ________________ % +/- 2 SD = ________________ % +/- 3 SD = ________________ %

68-95-99.7

What does an RR <1, =1, and > 1 mean

<1: event is less likely to occur in exposed =1: no difference in risk between two groups >1: event is more likely to occur in exposed

A randomized, double-blind, controlled trial is conducted to assess the efficacy of a drug to prevent osteoporotic fractures in women older than 65 years. Sixty women are enrolled; 30 subjects receive the new drug (group X), and 30 subjects receive a placebo (group Y). Two years later, seven subjects from group X have had fractures, and 15 subjects from group Y have had fractures (p=0.06). The investigators conclude that there is not a significant difference between the treatments, and the new agent is not recommended. Which of the following raises the most concern about the validity of this conclusion? A ) Low power B ) Low external validity C ) Use of placebo D ) Use of blinding E ) Inappropriate study population

A ) Low power The power of an experiment is the probability that it can detect a treatment effect, if it is present. Power and sample size go hand in hand and because of a low sample size, this experiment has low power and is therefore not effective in detecting a statistically significant difference But why not E?

An investigator conducts a study to assess the effectiveness of altering school lunches for improving serum lipid concentrations in middle school students. A low-fat diet is compared to the standard lunch. In order to minimize changes in the lunch habits of the students who will be given the low-fat diet (study group), the lower-fat lunch is given to five middle schools on the north side of the city. The control group consists of five schools on the south side of the city where the students will continue to receive the standard lunch. The results indicate that students in the study group have a significant decrease in their average serum cholesterol concentrations compared to the students in the control group who have an average increase in serum cholesterol concentrations. Which of the following raises the most concern about the validity of the study results? A) Biased measurement B) Lack of generalizability C) Lack of a placebo D) Lack of randomization E) Low power

A or D????

A cohort study is conducted to assess the relationship between oral contraceptive use and the risk of breast cancer. Participants are chosen from a closed-panel health maintenance organization (HMO] and include 530 women between 30 and 40 years of age with no gynecologic disease. Each participant is asked to complete a questionnaire about her use of an oral contraceptive. Twenty years later. 40% of the participants are enrolled in the same HMO. Of the 100 women who remained with the HMO and continued to use an oral contraceptive. 12 patients had developed histologically confirmed breast cancer compared with 4 patients among the women who did not use an oral contraceptive. Which ofthe following features of this study is most likely to affect its validity? A) Differential follow-up B) Lead time bias C) Misclassification of exposure D) Misclassification of outcome E) Recall bias

A) Differential follow-up Loss to follow up can cause bias if there is a greater tendency to loose subjects of a certain group, such as those taking oral contraceptives who developed breast cancer. In other words, there would be a differential loss to follow up with loss of diseased patients and this will skew the results Again, depending on which category is underreported as a result of differential loss to follow-up, either an underestimate or overestimate of effect (association) can occur.

Two programs for the treatment of patients with newly detected hypercholesterolemia were tried in a community. -Program A was used in one district of the community ,and Program b was used in another. -After 4 years ,40 % of the 110 patients on program A and 53 % of the patients on Program B had been successfully treated for hypercholesterolemia. -The P value for the difference was 0.3. -Based on this data,the health officials decide NOT to change to program B in the first district. Which of the following best explains this decision ? A-They attributed the difference in success rates to chance alone B-They distinguished between statistical significance and clinical significance C-They felt the power was too high to justify a decision in favor of program B D-They felt the p value was too small to justify a decision in favor of program B

A-They attributed the difference in success rates to chance alone P value of 0.3 is > 0.05 and suggests that 30% chance that the difference between the two is due to chance alone

Fill in the following *Power *Type 1 error *Type 2 error *alpha *beta *1-alpha A. B. C. D.

A. 1-alpha B. Type 1 error/ alpha C. Type 2 error/ beta D. Power (1-beta)

A study is conducted to assess relationship between age and serum cholesterol level in 3 groups. group X consist of 50 childrens (ages 6 to 10), group Y consists of 50 adolescents (ages 12 to 16), Group Z consist of 50 adults (aged 21-25). the serum total cholesterol level is measured in all participants. Which of the following tests is most appropriate to compare serum total cholesterol level among group X, Y and Z in this study? A: analysis of variance B: chi-square C: independent t-test D: multiple logistic regression E: paired t test F: Pearson correlation

A: analysis of variance Checks differences between means of 3 or more group

AR = the difference in risk attributable to the ___________ (exposure/treatment) ARR = the difference in risk attributable to the ___________ (exposure/treatment)

AR = exposure ARR = treatment

Accuracy vs Precision

Accuracy *Trueness of the test measurements (validity) *The absence of systematic error or bias in a test. *Systemic error reduces accuracy in a test A1.the consistancy and reproducibility of a test (reliability) Precision *The consistency and reproducibility of a test (reliability) *The absence of random variation in a test. *Random error reduces precision

Alpha ,Beta, and P Value

Alpha *probability of making a type 1 error * * Beta *probability of making a type 2 error * * P *Number between 0 and 1 *Determines if findings are due to chance alone *Compared to alpha /significance level. *If P < alpha, then statistically significant/ reject null *Most places classify statistically significant as P < 0.05 and statistically highly significant as P < 0.001 and therefore reject the null *P value of 0.05 = only 5% probability that findings were due to random chance

Define standard error of the mean What is the formula to find SEM?

An estimate of how much variability exists between the sample mean and the true population mean SEM = σ/√n

The scatter diagram shown records the height and weight of a random sample of male college students. The correlation (r) between height and weight is closest to which of the following? A) +1.00 B) +0.60 C) 0 D) -0.60 E) -1.00

B) +0.60 An r of 1/-1 would be a completely straight line of plots An r of near 1 is a strong positive correlation with few outliers

A study is conducted of patients with restrictive airway disease who have BMIs above normal. Results show a significant correlation (r = -0.25) between body fat and lung capacity. Which of the following is the most accurate conclusion based on these data? A) Increased body fat is associated with higher lung capacity B) Increased body fat is associated with lower lung capacity C) Twenty-five percent of individuals with increased body fat have a higher lung capacity D) Twenty-five percent of individuals with increased body fat have a lower lung capacity E) No relation exists between body fat and lung capacity

B) Increased body fat is associated with lower lung capacity -0.25 means negative correlation (as one variable inc, the other dec) E is not correct because for no relation to exist, r = 0

Five separate analyses for serum glucose are done using a new method on a single sample. The values obtained are 102,100, 96,101, and 98 mg/dL. These types of data are used to determine which of the following characteristics of the test? A) Accuracy B) Precision C) Predictive value D) Sensitivity E) Specificity

B) Precision We're trying a new method out and want to see how consistent and reproducible the results are. It's like checking your BP reading 5 times to make sure you get the same/similar result. You are checking for precision, not accuracy

Comparation of passing and failure rates at each of 3 test sites: A. T test B. Chi-square test C. One way anova. D. Two way anova E. Pearson correlation

B. Chi-square test Comparing categorical measures= Chi-square

Comparing the percentage of members of 3 different ethnic groups who have essential hypertension A. T test B. Chi-square test C. One way anova. D. Two way anova E. Pearson correlation

B. Chi-square test Comparing percentages = Chi-square

A paper says, "We chose the sample size to have an 80% power of detecting a 15% mean difference with a significance level (two-sided) of 5%." If there really is no difference between the groups overall, which of the following best represents the chance that the study will find a statistically significant difference, and what is this error called? A) 0%, Type I error 8) 0%, Type II error C) 5%, Type I error D) 5%, Type II error E) 15%, Type I error F) 15%, Type II error 6) 20%, Type I error H) 20%, Type II error

C) 5%, Type I error significance' level (two-sided) of 5% = p value and in this case we saw a difference when none really existed, so alfa error. C.

A study is designed to evaluate the relationship between ambient noise and hearing loss in an automobile manufacturing plant. Different locations in the manufacturing plant have strikingly different levels of ambient noise. Each employee is given a test for hearing acuity, and then the ambient noise level is measured at his or her workstation. Which of the following best describes this study design? A) Case series B) Cohort C) Crossover D) Cross-sectional E) Randomized clinical trial

C) Crossover

A program for the primary prevention of coronary artery disease is implemented in a community in the USA. Assuming that diagnostic procedures and detection remain the same, which of the following measures involving the disease is most effective in monitoring the program? A) Case fatality B) Hospitalization C) Incidence D) Mortality E) Prevalence

C) Incidence Primary prevention prevents new cases

A study is conducted to assess the effectiveness of a new blood test for early detection of prostate cancer. Ten thousand healthy men over the age of 50 years are randomly assigned to receive either annual rectal examination or annual screening with the new blood test. After 5 years, results show that of the 50 men in the blood test group that were diagnosed with prostate cancer, 40 were living 2 years after the diagnosis was made. In comparison, only 15 out of 45 men in the rectal examination group survived 2 years after being diagnosed with prostate cancer. Researchers conclude that the blood test increases survival compared with rectal examination. Which of the following potential flaws is most likely to invalidate this conclusion? A) Age of the patients B) Diagnostic bias C) Lead time bias D) Recall bias E) Type II error

C) Lead time bias Early detection makes it seem as though survival has increased, but the natural history of the disease has not changed

What is the formula for calculating Confidence Interval?

CI = mean +/- Z(SEM) SEM = standard error of the mean = σ/√n n = sample size For the 95% CI, Z = 1.96. For the 99% CI, Z = 2.58

Case controls measures what parameter? Cohorts measure what parameter?

Case control: Odds ratio Cohort: Relative risk Case control is comparing two groups with and without a disease. What is the odds of being exposed to a risk factor? Cohort is comparing two groups with and without a risk factor. What is likelihood of developing disease?

What is the purpose of a washout period in crossover studies?

Copy from UW

A study is conducted to compare 2 drugs (X&Y) for the treatment of arthritis pain. Fifty subjects are selected for the study.; Half are assigned to take drug X and half to take drug Y. One month later, pain relief is assessed. Following a 1 wk washout period (no drug), those who took drug X are reassigned o drug Y, and those who took drug Y, and those who took drug Y are reassigned to drug X. One month later, pain relief with two drugs is then compared across 5 subjects. Study design? Case control Case series Cross over Cross-sectional Historical cohort Prospective cohort Randomized clinical trial

Cross over Participants alternate receiving intervention and placebo Participants act as own controls

Two plans for the follow up treatment of newly detected hypertensive were tried in a community .plan A was used in the eastern and southern districts of the community while plan B was used in the northern and western districts. -In plan A the total number of hypertensives were 2200 and percent of hypertensives successfully treated were 41 % within 3 years. -In plan B the total number of hypertensives were 1900 and percent of hypertensives successfully treated were 45 % within 3 years. -The difference in sucess rates for the two plans were statistically significant (P < 0.01). -Health officials ,however ,decided not to change to plan B in the eastern and southern districts because the magnitude of the difference was so small.This implies.... A-They attributed the difference in success rates to chance alone B-They feel the samples were too small to justify a decision in favor of plan B. C-They felt the p value was too small to justify a decision in favor of plan B. D-They distinguished between the statistical significance and practical significance of the difference in success rates. E-They felt that the success rates should be higher than 50%

D-They distinguished between the statistical significance and practical significance of the difference in success rates. P < 0.01 and is therefore statistically significant, however health officials felt as if the difference in success rates of 4% was not enough to make a practical/clinical significance Because P < 0.01 and is therefore statistically significant, A is wrong. Sample sizes are fairly large so B is wrong

https://books.google.com/books?id=U_jLDQAAQBAJ&pg=PA46&lpg=PA46&dq=t-test+usmle&source=bl&ots=DyN0tXL3j_&sig=9yuW4Vt0rPHB8oPAFsW6JSWuCSE&hl=en&sa=X&ved=0ahUKEwjSo7ru59DUAhXJMSYKHbb3CB0Q6AEIdTAJ#v=onepage&q=t-test%20usmle&f=false

DO THE KAPLAN QUESTIONS DO ALL MEDBULLET QS DO ALL FIRECRACKER QS

http://www.medbullets.com/step1-stats/1001/types-of-studies

DO THESE QS

The prevalence of breast cancer is compared in two groups of women based on parity. The following data are obtained at age 70 in both groups .Children Yes No Breast cancer.....120......................180.........(300) No Breast ca......1380....................820..........(2200) .................................(1500)...................(1000).......(2500) Based on these data what is the Relative risk(Risk Ratio) for development of Breast cancer in childless women compared with women who have children? A-0.67 B-1.2 C-1.5 D-1.8 E-2.25

E-2.25 Asking for risk of unexposed/ risk of exposed exposure in this case = children risk of exposed = 120/1500 = 40/500=.08 risk of unexposed = 180/1000 = .18 RR = .18/.08 = 2.25

Two experimental drugs, drug X and Y, are being evaluated for the treatment of Congestive heart failure. patients receiving drug X have a cardiac index of 2.5 L/m2 with a 95% confidence interval of 1.5 to 3.5. patients receiving drug Y have a cardiac index of 1.7 L/m2 with a CI of 0.7 to 2.7. A test of the significance of difference shows a p-value of 0.1. which of the foll. is the likelihood that the diff in mean cardiac index of patients receiving drug X and drug Y is due to chance ? A. 0% B. 2.5% C. 5% D. 7.5% E. 10% F. 66.7% G. 95%

E. 10% P value of 0.1 = 10% chance that the difference between the two drugs was by chance alone

An investigator wishes to assess whether vaccination increases the incidence of developmental diseases such as autism. He asks parents of recently diagnosed cases about their child’s vaccination history and compares their responses to those of parents of children who have not been diagnosed with autism. What is the most significant type of potential bias present in this study design? A. Confounding bias B. Lead-time bias C. Length bias D. Measurement bias E. Recall bias

E. Recall bias Recall bias occurs when cases and controls remember information differently, ie, when a recent diagnosis may have prompted them to search their memory for recent events that they feel may have contributed to the diagnosis. A false association may be found because cases are more likely to remember recent vaccinations, not because of any pathological process.

Ecological study vs Cross-sectional study

Ecological study: Unit of analysis is populations Cross-sectional study: Unit of analysis is individuals

500 workers with bladder cancer and 200 without bladder cancer are selected for a study. History obtained of aniline dye exposure. Of the 500 with bladder cancer, 250 have exposure to aniline dye. Of the 200 without bladder cancer, 50 have exposure. What's the odds ratio for exposure variable? A 1/3 B 1/2 C 1.5 D 1 E 2 F 3

F 3 Odds ratio = odds that group with disease was exposed / odds that group without disease was exposed *Remember that for odds, we do NOT divide those who are/are not exposed to the total group with/without disease* 500 with bladder cancer, 250 exposed, 250 not exposed 200 without bladder cancer, 50 exposed, 150 not exposed OR = 250/250 / 50/150 = 1/ 1/3 =3

A study is conducted to compare the incidence of myocardial infarction in patients undergoing two different types of angioplasty or an operative procedure to manage single-vessel coronary artery disease. A total of 1000 patients are enrolled. Through a chance process, 500 are assigned to undergo the operative procedure, 250 are assigned to undergo one type of angioplasty, and 250 are assigned to undergo a second type of angioplasty. All patients are followed for 3 years to determine the incidence of myocardial infarction. Which of the following best describes this study design? A) Case-control study B) Community intervention trial C) Ecological study D) Historical cohort study E) Prospective cohort study F) Randomized clinical trial

F) Randomized clinical trial A randomized clinical trial is an experimental study while a prospective cohort study is an observational study Observational studies have no human interactions. Simply observe

A study is conducted to evaluate intelligence quotient (IQ) scores for patients with various types of schizophrenia. Twenty patients were dropped from the sample because they could not complete two or more portions of the test. Four additional patients refused to take the test. Results for the 100 patients who completed the test showed an average IQ of 110 with a standard deviation of 20. Which of the following is the best estimate of the 95% confidence interval for this sample? A. 70 to 130 B. 70 to 150 C. 85 to 115 D. 90 to 130 E. 105 to 115 F. 106 to 114

F. 106 to 114 CI = mean +/- Z(SEM) = 110 +/- 2 (20/10) =110 +/-4 =106 to 114

A new antiplatelet agent is developed for the prevention of recurrence of stroke. In a large randomized clinical trial with equal numbers of men and women, the rates of stroke are lower in patients receiving the new agent than in patients receiving the standard treatment. Results are shown: Recurrent Stroke Rates per 1000 Person-Years Standard Treatment New Antiplatelet Drug Women .12 .04 Men .24 .08 Overall .18 .06 Based on these results, which of the following is the relative risk reduction in women? A) 8% B) 12% C) 16% D) 33% E) 50% F) 62% G) 67% H) 75%

G) 67% RR = .04/.12 = .33 RRR = 1-RR = .67

In a pharmaceutical study, Group A is the placebo group and Group B is the group that received the new drug. Data were gathered on Groups A and B and confidence intervals were calculated. Side effect rates were calculated as a percentage of each group. Which of the following grouping comparison statistics are significantly different? A) Group A confidence interval 30%-46% and Group B confidence interval 44%-88%. B) Group A confidence interval 10%-30% and Group B confidence interval 44%-88%. C) Group A confidence interval 0.1%-0.3% and Group B confidence interval 0.1%-0.3%. D) Group A confidence interval 88%-90% and Group B confidence interval 88%-90%. E) None of the above is statistically significant

Group B confidence interval 44%-88%. *If the CIs between 2 groups do not overlap, a statistically signifcant difference exists.

Tests with high sensitivity have a low rate of false ___________ (negative/positive) Tests with high specificity have a low rate of false ___________ (negative/positive)

High sensitivity: low false negative High specificity: low false positive (the respective denominators in both formulas)

Define standard deviation What is the formulat to find SD?

How much variability exists from the mean in a set of values SD =

Internal vs External Validity

Internal validity Are the results of the study valid for the population of who where actually studied ? External validity Are the results of the study valid for other patients

Which of the measures of central tendency is most affected by outliers? least?

Most: Mean Least: Mode

Compare the formulas for NNT and NNH

NNT = 1/ARR NNH = 1/AR

How are relative risk and odds ratio related?

OR overestimates RR when disease is common (makes the exposure look even more bad when it is harmful and makes the exposure look even more good when it is beneficial) OR closer to RR when disease is rare

A case-control study is conducted to assess the possible relationship between regular use of aspirin and colon cancer. The odds ratio for colon cancer among aspirin users is 0.6 with a 95% confidence interval of 0.4 to 0.9. Which of me following is the most accurate conclusion? A) The attributable risk of colon cancer from aspirin use is 6 per 1000 person-years B) The incidence of colon cancer among aspirin users is 6 per 1000 person-years C) The incidence of colon cancer among nonusers of aspirin is 6 per 1000 person-years D) The probability of confounding bias is less than 5% E) The risk of colon cancer is 40% lower in aspirin users than nonusers

Odd's ratio = odds that the group with the disease was exposed to risk / odds that the group without the disease was exposed = 0.6 = 6/10 E) The risk of colon cancer is 40% lower in aspirin users than nonusers

What is the formula for PPV What is the formula for NPV?

PPV = TP/All positive NPV = TN/All negatives

How does prevalence/pretest value affect the positive and negative predictive value?

Positive predictive value is directly correlated with prevalence. The higher the pretest value/prevalence, the higher the value of the positive predictive value (more likely that a positive is a true positive) Negative predictive value is inversely correlated with prevalence. The higher the pretest value/prevalence, the lower the value of the negative predictive value (less likely that a negative is a true negative)

In a positively skewed distribution, the tail trails off to the __________ (right/left) In a negatively skewed distribution, the tail trails off to the __________ (right/left)

Positively skewed: Right Negatively skewed: Left

Primary vs Secondary vs Tertiary Prevention

Primary **Prevent* a disease from occuring *ex: health promotion: regular exercise, weight loss Secondary **Detecting* a disease before it causes symptoms *ex: screening methods (BP tests, Pap smears) Tertiary **Treating* a disease to prevent progressions/complications *ex: blood sugar/BP control in diabetes

Standard 2-2 Contingency table for evaluation of diagnostic tests

Proportion of all people with disease who test positive, or the probability that when the disease is present, the test is positive

What does the specificity of a test describe?

Proportion of all people without disease who test negative, or the probability that when the disease is absent, the test is negative

Define Negative Predictive Value

Proportion of negative test results that are true negative. Probability that a person with a negative test result actually does not have the disease

Define Positive Predictive Value

Proportion of positive test results that are true positive. Probability that a person who has a positive test result actually has the disease

Compare the formulas for RR and AR

RR = a/(a + b)/c/(c + d AR = a/(a + b) - c/(c + d

The prospected study was conducted to see the selinium ingestion with the gastric cancer. The investigator estimated the Relative Risk of 0.3 who were ingested the High Selenium diet (95% confident Interval 0.1-0.8) which of the following is the most approriate interpretation A. Selenium cause Gastric Ca B. Seleniun does not cause gastric Ca. C Selenium is higher risk of Gastric ca D. selenium is low risk of Gastic Ca E. selenium is positively correlated with Gastric Ca

RR = risk of exposed/risk of unexposed = 0.3 = 3/10 RR < 1 means the event is less likely to occur in exposed.

What is the formula for specificity

TN/TN + FP

What is the formula for sensitity?

TP/TP + FN

Type 1 error is also known as a False ____________ (positive/negative) Type 2 error is also known as a False ____________ (positive/negative)

Type 1: False positive Type 2: False negative

What does the sensitivity of a test describe?

Use FA def

Example case of confidence interval

We measure the heights of 40 randomly chosen men, and get a: mean height of 175 cm, SD of 20 cm 95% CI = between 168.8cm and 181.2cm This means that the true mean of ALL men (if we could measure their heights) is likely to be between 168.8cm and 181.2cm. However there is a 5% chance that the true mean will not fall within this confidence interval

A large study of serum folate levels in women aged 16-45 years reveals that this parameter is normally distributed, with a mean of 5 ng/ml and a SD of 0.5 ng/ml. According to the study results, 95% of serum folate observations in these patients will lie between the following? a- 4 and 6 b- 4 and 5.5 c- 4.5 and 5.5 d- 3.5 and 6 e- 3.5 and 6.5

a- 4 and 6 95% = 2SD from mean

A randomized clinical trial is conducted to compare wound healing and cosmetic differences between two surgical procedures for closing skin wounds following cesarean delivery. A total of 500 women undergoing c-section during a 6-month period are enrolled in the study. A research asst is assigned to conduct outcome assessments of each participant at 2 wks and 6 wks following delivery. both procedures require the expertise of surgical specialists who cannot be blinded to the procedure. results show that procedure A has a lower rate of wound infection compared to procedure B (RR of 0.66 with a 95% CI of 0.30 - 1.45); The two procedures also have a similar cosmetic results 6 wks after delivery. Which of the following statements most accurately represents the comparision between procedure A and procedure B in a clinical care setting? a. Neither procedure is superior b. Procedure A is superior to procedure B c. Procedure B is superior to procedure A d. The two procedure should not be used in a clinical care setting

a. Neither procedure is superior When looking at confidence intervals for relative risk, relative benefit, etc, remember that if the CI 95% crosses 1, there is no difference between the groups

In a Case control study, 100men with bladder cancer are compared to 100 controls. The odds ratio of the study was 3. 95% confidence interval was 1.8-6.6 for association of bladder cancer to EtOH consumption. Interpretation? a. The likelihood is 95% that the true value for the odds ratio is between 1.8-6.6. b. There is a 95% chance that the true odds ratio is 3.00 c. At least 5% of consumers of alcohol will develop cancer of the bladder. d. Men who drink alcohol are 5 times more likely to develop bladder cancer than those who do not drink alcohol

a. The likelihood is 95% that the true value for the odds ratio is between 1.8-6.6. frg v n b gg

What is the formula to calculate Likelihood Ratio

a/a+c // b/b+d or TP/number of patients with disease / FP/number of patients without the disease or SN/1-SP

A 48 year old man presents with possible hypertension. Based on 20 measurements, his average diastolic pressure is 94 mmHg, SD is 8mmHg. If only four measurements were made rather than 20, which of the following statements would best describe the width of the 95% CI with regard to the mean blood pressure? a) Smaller b) Larger c) The same d) Changed, but the direction cannot be predicted

b) Larger CI = mean +/- Z(SEM) SEM = standard error of the mean = σ/√n if n decreases, you have a smaller denominator and SEM increases, thus Cl increases Another way to approach this problem is that the width of a confidence interval represents our uncertainty concerning the estimate. If we have a larger sample, we should be less uncertain about our estimate (narrower interval). Similarly, if we have a smaller sample, we will be more uncertain about our estimate (again, assuming that the only change is sample size).

In a study, certain test was found to always have different test-retest values. Which of the following can be concluded? a) Random error resulted in low accuracy of the test b) Random error resulted in low precision of the test c) Random error resulted in low validity of the test d) Systematic error resulted in low accuracy of the test e) Systematic error resulted in low precision of the test f) Systematic error resulted in low validity of the test

b) Random error resulted in low precision of the test

When describing the characteristics of a study, what is the difference between the n and the N of the study?

n = sample size N = population size

t-test compares means of _______ groups ANOVA compares means of _______ groups

t-test: 2 (tea for 2) ANOVA: 3 (analysis of variance = 3 words)