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;
109 Cards in this Set
- Front
- Back
Incidence density |
number of new cases of a diseaseoccurring in a specified time period . Total amount of "person-time" at riskcontributed during the time period |
|
Incidence density |
Estimates the instantaneous rate ofoccurrence of disease per unit of timerelative to the size of the population at risk |
|
Refers to theprevalence measured for a specific timeinterval |
Period prevalence |
|
Number of deaths= due to a disease in a specified time period /total number of deaths during that time period |
Proportional mortality (Can be misleading if mortality rates for othercauses are unusually high or low in a group) |
|
Relationship between prevalence, incidence, and time? |
incidence X duration *(when prevalence is low, i.e.,<10%) Holds only when incidence and duration are stable overtime. Useful in predicting what a change in one variable willcause in another |
|
Relationship among mortality, incidence, and case fatality? |
mortality = incidence X case fatality (when incidence and case fatality arestable over time) |
|
Relationship between survival rate and case fatality |
1-case fatality rate |
|
How is birth rate calculated? |
Number of live births in a year .Population (in thousands) at midyear |
|
How is fertility rate calculated? |
number of live births reported in one yearnumber (in thousands) of womenage 15-44 years at midyear |
|
How is the fetal death rate (stillbirth rate) calculated? |
annual number of fetal deaths (gest. age 20 wks/350 g) /annual number of fetal deaths plus live births(in thousands) |
|
How is the neonatal death rate calculated? |
annual number of deaths in the first 28 days of life /annual number of live births (in thousands) |
|
How is infant death rate calculated? |
annual number of deaths in the first year of life/ annual number of live births (in thousands) |
|
Takes into account not only the cause ofdeath but the age of occurrence? |
Years of potential life lost |
|
How is YPLL calculated? |
Calculated by multiplying the number ofcause-specific deaths in an age groupand multiplying by the difference betweenthe midpoint of the age group and age 75(or the average age at death) |
|
"Cohort effect" |
a point-in-time, cross sectionalobservation that reflects variationin disease rates based on year of birth, andvariation in disease rate within each cohortby age Suspect whenever you see an unexpecteddecline in older age groups |
|
How is relative risk calculated? |
a (a+ b) / (c + d) incidence in the exposed/ incidence in the unexposed |
|
Attributable risk |
Accounts for the baseline incidence of diseaseand gives the absolute amount of excess risk anindividual |
|
True or false? Attributable risk can be calculated from case-control studies? |
False, requires a prospective study |
|
Calculation for NNT? |
1/Attributable risk |
|
Attributable risk percent (AR%) = |
incidence in exposed - incidence inunexposed unexposed / incidence in exposed |
|
What is the PAR%? |
the amount of disease thatwould be prevented if the risk factorcould be eliminated from the population |
|
How is PAR% calculated? |
Total incidence - incidence in the unexposed / total incidence |
|
True or false? PAR% can also be calculated from casecontrolstudies using the odds ratio? |
True |
|
What are three types of bias? |
* Measurement bias * Recall Bias * Selection bias |
|
What is the ecologic fallacy? |
the individuals with theexposure aren’t necessarily the oneswith the outcome |
|
What is a major disadvantage in matching in case-control studies? |
the individuals with theexposure aren’t necessarily the oneswith the outcome |
|
Occurs when a phenomenon is more likely to be observed for a particular set of study subjects. For instance, the syndemic involving obesity and diabetes may mean doctors are more likely to look for diabetes in obese patients than in thinner patients, leading to an inflation in diabetes among obese patients because of skewed detection efforts. |
Detection bias |
|
Refers to the phenomenon that the performance of a diagnostic test may vary in different clinical settings because each setting has a different mix of patients. Because the performance may be dependent on the mix of patients, performance at one clinic may not be predictive of performance at another clinic. |
Spectrum bias if it is a case-control design where a healthy population ('fittest of the fit') is compared with a population with advanced disease ('sickest of the sick'); that is two extreme populations are compared, rather than typical healthy and diseased populations. |
|
Arises when the researcher unconsciously influences the experiment due to cognitive bias where judgement may alter how an experiment is carried out / how results are recorded. |
Observer bias |
|
Information bias» |
Misclassification (including heterogeneousoutcomes)» Differential reporting of exposure data (includingrecall bias) |
|
Misclasification and odds ratios |
|
|
What best describes this study design: the exposed and unexposed are reviewed retrospectively to see who developed the disease? |
|
|
When should a case-control study be conducted? |
* Little known about the disease * Many risk factors could contribute to outcome * Rare disease * Other designs too expensive * long latent period between exposure and outcome. |
|
True or false? Case selection in case control studies should rely on prevalent cases. |
False. Incident cases are preferred. |
|
True or False of case control studies? Over-sampling of cases of long duration maytend to bias the results, describing factors thatinfluence prognosis or survival rather thanetiologic factors. |
True |
|
Rule of thumb for selecting controls in case -control studies |
Choose controlsubjects who, if they had gotten thedisease under study, would have beeneligible for case selection |
|
How are case -control studies assembled? |
|
|
How are cohort (pro/retrospective) studies assembled? |
|
|
What are some criteria for using cohort studies? |
* Risk factor represents a rare event * Intent to study the multiple potentialoutcomes of a single exposure * Necessary if incidence rates are neededfrom the study *Necessary if limitations make otherdesigns unfeasible |
|
This study type outcome measures include relative risk,attributable risk, and population attributablerisk percent |
Cohort study |
|
What is a major type of bias in cohort studies? |
Ascertainment bias can occur,especially if method of determiningoutcome does not include blinding toexposure status |
|
Major problems with retrospective cohort studies? |
*Lost to follow-up * Sample sizes are usually smaller * Misclassification of either exposure oroutcome is potentially a greater problem * Data on exposure status for potentialconfounders may not be available |
|
Most often used when a potential confounder isidentified in the analysis as an importantdeterminate of excess disease risk |
|
|
What best describes intention to treat analysis? |
subjects must be analyzed as belongingto the group to which they were first randomized,even if they cross-over to the other group |
|
What are two major concerns with quasi-experimental studies ? |
Additional sources of bias not controlled for byrandom assignment (especially selection bias)» interclass correlation that may be responsiblefor the observed effect, unrelated to theintervention (confounding) |
|
Types of quasi-experimental studies ? |
before/after, Non-equivalent control group design, |
|
Study groups are assembled in a nonrandomizedfashion intended to minimizeunequal distribution of important confounders,and researcher decides which group(s) getsthe intervention |
Non-equivalent control group design, Time series design |
|
A sequence of data points, typically consisting of successive measurements made over a time interval. Examples of time series are ocean tides, counts of sunspots, and the daily closing value of the Dow Jones Industrial Average. |
Time series analysis |
|
Time series graph |
|
|
A kind of time series analysis, where units areassigned to a condition based on a cutoffscore on a measured covariate; for example,for a smoking cessation intervention,communities that exceed a certain cutoff forpacks of cigarettes sold are given theintervention, and communities below thatcutoff are the comparison |
Regression Discontinuity |
|
|
Regression discontinuity example |
|
Requirement for a factor to be a confounder? |
The potential confounder must beassociated with the outcome of interest * the confounder is an actual risk factor forthe outcome * the confounder affects the likelihood ofrecognizing the outcome The potential confounder must beassociated with the exposure of interestbut not be a result of the exposure |
|
What are the three confounding through study design? |
1) Restriction: study only the subjects in a given category 2) Matching 3) RCT |
|
When is something not a confounder? |
*If an individual's status regarding the confounder isa result of the exposure under study, or theconfounder is in the "causal pathway" betweenexposure and outcome *If an individual's status regarding the confounder isa result of the disease under study * If the confounder is essentially measuring thesame thing as the exposure * If the association between the confounder and theoutcome of interest is thought to be due to chance |
|
You essentially estimatehow the comparison would turn out if thegroups had the same distribution of theconfounder (e.g., the same age distribution) |
Rate adjustment is most often used toaccount for the affect of age (ageadjustment); also commonly used to adjustfor gender; could be used to adjust for anyconfounder or group of confounders |
|
The standardized mortality ratio |
Observed number of deaths in study group/Expected number of deaths in study group |
|
interpreted as the ratio of the observed number of events in each comparison group to the number that would be expected based on the event rates from the standard population |
The standardized mortality ratio |
|
True or false? indirect adjustment is commonly used toprovide adjusted rates for descriptivepurposes, and in comparing data fromseveral different studies |
False: Direct adjustment commonly used to provide adjusted rates for descriptive purposes, and in comparing data from several different studies |
|
True or false? Indirect adjustment may be preferablewhen there are only a small number ofpeople in each age category of thestudy groups where the age-specificrates are based on small numbers andchanges of only one or two individualsin the numerator produce big changesin the age-specific rates |
True |
|
describes how often ascreening test detects a disease when itis indeed present; = TP/(TP+FN), or truepositives over the total with disease |
Sensitivity |
|
Specificity |
describes how often ascreening test detects the absence ofdisease when it is indeed absent; =TN/(TN+FP), or true negatives over thetotal without disease |
|
describes howoften individuals with positive testsactually have the disease; =TP/(TP+FP), or true positives over allpositives Negative predictive value describes |
Positive predictive value |
|
describes howoften individuals with negative tests areactually disease-free; = TN/(TN+FN), ortrue negatives over all negatives |
Negative predictive value |
|
If fecal occult blood testing is 92%sensitive and 95% specific, and theprevalence of colon cancer is 2/1000,what is the positive predictive value of apositive test (i.e., what percent ofpositives will have colon cancer)? |
|
|
True of false as specificity increase, positive predictive value increases |
True |
|
In this type of testing, the overall screening result ispositive if any one test is positive; this strategyincreases sensitivity at the expense of specificity |
Parallel testing |
|
In this type of screening method, the overall screening result ispositive only if all tests are positive; this strategyincreases specificity at the expense of sensitivity |
Series testing (sequential testing) |
|
Agent infects hostcontinuously with no overt evidence ofdisease or infection |
Colonization |
|
Agent infects host,time-limited, with no overt indication(majority of infections) |
Covert infection |
|
Infection with disease |
Overt infection (minority of disease) |
|
Important in diseasetransmission and control, and influencesapparent epidemiology (understates amount ofdisease and overstates severity) Clinical disease frequent, few severe cases Infection usually always fatal |
Inapparent infection frequent, rare clinicaldisease |
|
Infectivity |
ability to involve/multiply ina host (establish an infection) |
|
ID50 |
dose of agent necessary to infect/kill50% of hosts |
|
Ability to be transmittedto other hosts |
Infectivity |
|
Ability to cause diseasein a susceptible host; to produceclinical illness |
Pathogenicity |
|
ability to produce severeclinical illness, including death |
Virulence |
|
ability to elicit animmune response |
Immunigencity (gain be positive or negative [e.g. drug antibodies) |
|
period between infection andmaximal communicability (can be the same asincubation time) |
Generation time |
|
measures the rate ofspread of disease within an exposed group = (#new cases - # initial cases)/(#susceptible - # initial cases) |
Secondary attack rate |
|
gradual increase in the frequency ofdisease occurrence above endemic level |
Hyperendemic |
|
a sudden increase in the frequency ofdisease occurrence above endemic level (clearly inexcess of expected levels) |
epidemic |
|
|
Classic epicurve |
|
Vaccine effaicacy |
(Attack rate (AR) in unvaccinated – AR in vaccinated/ AR in unvaccinated x 100 |
|
What are hills causality criteria?? |
1) consistency 2) strength of association 3) specificity : only one factor is consistently implicated 4) temporal association 5) coherence of explanation : fits with existing observations and makes logical sense 6) biological plausible 7) experiment: evidence from double blind RCT 8)analogy: relate to other causal associations if associations are similar 9) dose-response. However, absence does not exclude causality |
|
Name that survey: indiviuals within households questioned about daily activity limitations, acute and chronic diseases, physician or hospital visits |
National health interview survey |
|
What are hills causality criteria?? |
1) consistency 2) strength of association 3) specificity : only one factor is consistently implicated 4) temporal association 5) coherence of explanation : fits with existing observations and makes logical sense 6) biological plausible 7) experiment: evidence from double blind RCT 8)analogy: relate to other causal associations if associations are similar 9) dose-response. However, absence does not exclude causality |
|
Name that survey: indiviuals within households questioned about daily activity limitations, acute and chronic diseases, physician or hospital visits |
National health interview survey |
|
What are some control measures during an outbreak? |
Containment, quarantine, sanitation, isolation, and prophylaxis . Diagnosis and treatment; control of vectors |
|
Front (Term) |
Zero growth |
|
Front (Term) |
Zero growth |
|
Front (Term) |
Slow growth |
|
Front (Term) |
Publication bias |
|
Front (Term) |
Life expectancy at birth Note that the life expectancy of a black female is about equal to that of a white male. The largest difference is between a white female and a black male. |
|
True or false smoking is a risk factor for prostate cancer incidence |
False. It is a risk factor for prostate cancer mortality |
|
Describe the incidence and death rates for lung cancer comparing African men and women to whites? |
Black Men-incidence about 20% higher; 11% lower for black women Incidence rates decreasing for men of both races; women of both races have stable incidence Death rates declined faster for AM than whites (in younger adults rates =); Cancer death rates are lower for AM women compared to white women; AA men have higher death rate. |
|
How do the incidence and death rates for cervical cancer compare for white and black women? |
Up to age 50 incidence rates = Then black > white
Death rates 2 x blacks women |
|
How does the incidence and death rates compare for blacks vs whites |
A significant disparity exists between blacks and whites in both incidence and mortality. |
|
True or false? The incidence and mortality rate of prostate cancer is 63% and 2.3 xx higher respectively? |
True |
|
. You are studying the effect of a new vaccine on the incidence of malaria in 2,000 malaria-naive subjects. 1,000 subjects were randomly assigned to receive the vaccine and 1,000 were randomly assigned to receive a placebo. All subjects were followed for 1 year. At the end of the trial, you find that 20 new cases of malaria occurred in the vaccine group and 40 new cases occurred in the placebo group. You also find that the average censored time was 40 weeks in the vaccine group and 52 weeks in the placebo group. Which of the following can be concluded from these data A) New cases of malaria occurred sooner on average in the vaccine group compared with the placebo group. B) New cases of malaria occurred later on average in the vaccine group compared with the placebo group. Correct answerC) More subjects dropped out of the vaccine group compared with the placebo group. D) More subjects dropped out of the placebo group compared with the vaccine group. |
Answer: C. In a longitudinal study, a subject is censored when the drop out of the trial or if they complete the trial without having the outcome of interest. Because the follow-up time of this trial is 1 year, the expected mean censoring time would be 52 weeks if all of the subjects stayed in the trial. The fact that the mean censoring time in the vaccine group was less than 52 weeks indicates that there were subjects who dropped out of the trial early in that group |
|
A study examining the association between ultraviolet B exposure and incidence of colorectal cancer uses linear regression to assess the relationship between latitude and national incidence rates of colorectal cancer, controlling for average cloud cover. This study is best described as which type of study? A) Case-control study B) Cohort study C) Cross-sectional study Correct answerD) Ecological study |
Answer: D. The data that are the basis for this study are group data (exposure by geographic location and outcome in national incidence rate), which defines an ecologic study, rather than individual data, which are used in the other study designs (i.e., exposure measured for each individual and linked to outcome in that individual (disease yes/no). |
|
85. You conduct a study comparing a new method for detecting atherosclerotic plaques in coronary plaques with an old method in 100 patients with known coronary artery disease. The sample size was determined by setting the alpha error (two-sided) to 0.05 and the beta error to 80% for the effect size determined to be clinically meaningful. You find that the new method detects 50% more of the plaques that are confirmed by angiography (the gold standard) than the old method, with a 95% confidence interval of the difference of (-25%, 75%). What is the best conclusion from this study?
The new method is not superior to the old method. Missed correct answerB) The study power is low. C) The alpha error is high. D) The study was subject to observer bias. |
Answer: B. Power is equal to 1-beta, where the beta error is the probability of accepting a null hypothesis when it is actually false. In this case, the beta error is 80% (with a corresponding power of 20%), meaning there is an 80% chance of not finding a significant difference that actually exists in the study. This is a high chance of making this Type II error, corresponding to a low power, and invalidating the interpretation of no difference (A.). The alpha error is the probability of rejecting a null hypothesis that is actually true. In this case, the alpha error was set at 5%, (p<.05), which is standard for most studies. The 95% confidence interval includes 0 in this study, which is comparing differences (expressed as a percentage), so p>.05. A lower alpha (e.g., .01) would make it even less likely to find a statistically significant difference. Thus C. is incorrect. There is no information in the problem to indicate observer bias (D.). |
|
On average male smokers die how many years earlier than male nonsmokers?
A) 5 years B) 8 years Missed correct answerC) 13 years Incorrect answerD) 18 years Points: 0 out of 1 Feedback:
|
A) 5 years B) 8 years Missed correct answerC) 13 years Incorrect answerD) 18 years Points: 0 out of 1 Feedback:
Answer: C. Actuarial data show that male smokers die an average of 13 years earlier than their non-smoking counterparts; for females, the rate is about 15 years earlier for smokers compared with non-smokers.
|
|
. You are studying the effect of a new vaccine on the incidence of malaria in 2,000 malaria-naive subjects. 1,000 subjects were randomly assigned to receive the vaccine and 1,000 were randomly assigned to receive a placebo. All subjects were followed for 1 year. At the end of the trial, you find that 20 new cases of malaria occurred in the vaccine group and 40 new cases occurred in the placebo group. You also find that the average censored time was 40 weeks in the vaccine group and 52 weeks in the placebo group. Which of the following can be concluded from these data A) New cases of malaria occurred sooner on average in the vaccine group compared with the placebo group. B) New cases of malaria occurred later on average in the vaccine group compared with the placebo group. Correct answerC) More subjects dropped out of the vaccine group compared with the placebo group. D) More subjects dropped out of the placebo group compared with the vaccine group. |
Answer: C. In a longitudinal study, a subject is censored when the drop out of the trial or if they complete the trial without having the outcome of interest. Because the follow-up time of this trial is 1 year, the expected mean censoring time would be 52 weeks if all of the subjects stayed in the trial. The fact that the mean censoring time in the vaccine group was less than 52 weeks indicates that there were subjects who dropped out of the trial early in that group |
|
A study examining the association between ultraviolet B exposure and incidence of colorectal cancer uses linear regression to assess the relationship between latitude and national incidence rates of colorectal cancer, controlling for average cloud cover. This study is best described as which type of study? A) Case-control study B) Cohort study C) Cross-sectional study Correct answerD) Ecological study |
Answer: D. The data that are the basis for this study are group data (exposure by geographic location and outcome in national incidence rate), which defines an ecologic study, rather than individual data, which are used in the other study designs (i.e., exposure measured for each individual and linked to outcome in that individual (disease yes/no). |
|
85. You conduct a study comparing a new method for detecting atherosclerotic plaques in coronary plaques with an old method in 100 patients with known coronary artery disease. The sample size was determined by setting the alpha error (two-sided) to 0.05 and the beta error to 80% for the effect size determined to be clinically meaningful. You find that the new method detects 50% more of the plaques that are confirmed by angiography (the gold standard) than the old method, with a 95% confidence interval of the difference of (-25%, 75%). What is the best conclusion from this study?
The new method is not superior to the old method. Missed correct answerB) The study power is low. C) The alpha error is high. D) The study was subject to observer bias. |
Answer: B. Power is equal to 1-beta, where the beta error is the probability of accepting a null hypothesis when it is actually false. In this case, the beta error is 80% (with a corresponding power of 20%), meaning there is an 80% chance of not finding a significant difference that actually exists in the study. This is a high chance of making this Type II error, corresponding to a low power, and invalidating the interpretation of no difference (A.). The alpha error is the probability of rejecting a null hypothesis that is actually true. In this case, the alpha error was set at 5%, (p<.05), which is standard for most studies. The 95% confidence interval includes 0 in this study, which is comparing differences (expressed as a percentage), so p>.05. A lower alpha (e.g., .01) would make it even less likely to find a statistically significant difference. Thus C. is incorrect. There is no information in the problem to indicate observer bias (D.). |
|
On average male smokers die how many years earlier than male nonsmokers?
A) 5 years B) 8 years Missed correct answerC) 13 years Incorrect answerD) 18 years Points: 0 out of 1 Feedback:
|
A) 5 years B) 8 years Missed correct answerC) 13 years Incorrect answerD) 18 years Points: 0 out of 1 Feedback:
Answer: C. Actuarial data show that male smokers die an average of 13 years earlier than their non-smoking counterparts; for females, the rate is about 15 years earlier for smokers compared with non-smokers.
|
|
Which of the following has NOT been shown to be a risk factor for chronic obstructive pulmonary disease (COPD)A) Air pollution B) Occupational dusts C) Passive tobacco smoke exposure Correct answerD) Recurrent pneumonia |
Answer: D. The most significant risk factor for COPD is exposure to tobacco smoke, primarily one's own smoking, but also large amounts of second-hand smoke. Long-term exposure to occupational dusts and chemicals and air pollution are also major risk factors for the disease. Recurrent pneumonia is a complication of COPD, not a risk factor |
|
Front (Term) |
Years of potential life lost calculation |
|
Match the following: A) BRFSS - run by CDC and state governments B) National Health Interview survey: designed by the CDC national health statistics branch; administered by the US census Bureau US census: survey mailed first NHANES: National Center for health statistics: National health care statistics runs it. Some topics include quality (e.g., pt safety), health disaparities and disaster preparedness |
All are matched correctly. |