Epidemiology and biostats Flashcards
In what kind of study is Relative risk is used?
Cohort
In what kind of study is odds ratio is used?
Case-control
Interpretation of RR and OR?
> 1: positive association
= 1: no association
< 1: negative association (protective)
What is sensitivity
Given that I have the disease how likely is that I’ll have a positive test
True positive / Disease (+) = True positive / (True positive + False negative)
Sensible tests are used as screening tests
What is specificity
Given that I am disease-free how likely is that I’ll have a negative test
True negative / Disease (-) = True negative / (False positive + True negative)
Specific tests are used as a confirmatory test after a positive screening
What is positive predictive value
Given that I have a positive test how likely is that I have the disease
True positive / Positive test = True positive / (True positive + False positive)
The higher the prevalence, the higher the PPV will be
What is negative predictive value
Given that I have a negative test how likely is that I don’t have the disease
True negative / Negative test = True negative / (True negative + False negative)
Accuracy vs precision
Precision measures the consistency of the results (if I repeat the test, how likely is that I will have the same results)
Accuracy measures that the test measure what it is intended to measure
Bias from participants and how to address them
- Hawthorne effect: People change their behaviour because they know they’re being observed
- Recall bias: sick patients remember more
Addressed by blinding
Bias from researcher and how to address them
Selection bias: at baseline, the groups are different in way they shouldn’t be
• Addressed by randomization and matching
Observer bias: The researcher knows who is in control and intervention groups
• Addressed by blinding
You find in a study that ice-cream consumption is associated with drowning.
What type of bias is this?
Confounding:
Relationship between exposure and outcome is distorted because a third factor is related to both exposure and outcome.
In this case the third factor is summer
OCP have a small association with DVT. However, when smoking, the association between OCP and DVT is highly enhanced.
Type of bias?
Effect modification:
The relationship between exposure and outcome is enhanced by a third factor, which only affects the outcome and not the exposure
Case-control study or cohort study. Which one is better for a rare disease?
Case-control study
What is the type I error?
Saying there’s a difference when in reality there is none
What is the type II error?
Saying there’s no difference when in reality there is
What is primary prevention?
o They don’t have yet the disease
o Goal: Keep the patient healthy
o Reduce exposure to risk factors
o Examples: weight loss, smoking cessation, reduce EtOH, healthy eating. vaccination
What is secondary prevention?
o They already have the disease
o Goal: delay progression by detecting the disease in an early stage (screening) and start early treatment if needed
What is tertiary prevention?
o Goal: prevent complications from an existing disease with acting therapy
Colon cancer screening guidelines
Start: 50 yrs or 10 yrs before the Dx of a primary relative
How: Flex sigmoidoscopy / 10 yrs Or Fecal occult blood test / 2 yrs
Stop: 75 yrs
Breast cancer screening guidelines
Start: 50 yrs
How: Mamography / 2 yrs
Stop: 75 yrs
Cervical cancer screening guidelines
Start: 21 yrs
How: Pap smear / 3 yrs
Stop: 70 yrs if 3 consecutive negative pap smears
Lungcancer screening guidelines
Start: 55–75 with history of > 30 pack-year who quit < 15 yrs ago
How: Low dose CT scan / yrs
Stop: 80 yrs or quit > 15 yrs ago
Cancers you don’t screen for
Prostate and ovarian
Abdominal aortic aneurism. Who and how to screen?
Men > 65 and Women > 65 who have ever smoked
Abdominal U/S
Osteoporosis. Who and how to screen?
Women > 65
DEXA scan (if positive, bisphosphonates)
Hep C. Who and how to screen?
Baby boomer (1945–1965)
Hep C Ab
HIV. Who and how to screen?
Everyone
ELISA
HTN, Who and how to screen?
Everyone
Ambulatory monitoring
DM. Who and how to screen?
> 40 or at high risk / 3 yrs
A1C
Dyslipidemia. Who and how to screen?
♀ 45 yrs; 30 if high risk
♂ 35 yrs ; 25 if high risk
Lipid panel
Bias introduced into a study when a clinician is aware of the patient’s treatment type.
Observational bias.
Bias introduced when screening detects a disease earlier and thus lengthens the time from diagnosis to death.
Lead-time bias.
If you want to know if geographical location affects infant mortality rate but most variation in infant mortality is predicted by socioeconomic status, then socioeconomic status is a _____.
Confounding variable.
The number of true positives divided by the number of patients with the disease is _____.
Sensitivity.
Sensitive tests have few false negatives and are used to rule _____ a disease.
Out.
PPD reactivity is used as a screening test because most people with TB (except those who are anergic) will have a (+) PPD. Highly sensitive or specific?
Highly sensitive for TB.
Chronic diseases such as SLE—higher prevalence or
incidence?
Higher prevalence.
Epidemics such as influenza—higher prevalence or
incidence?
Higher incidence.
Cross-sectional survey—incidence or prevalence?
Prevalence.
Cohort study—incidence or prevalence?
Incidence and prevalence.
Case-control study—incidence or prevalence?
Neither.
Describe a test that consistently gives identical results, but the results are wrong.
High reliability, low validity.
Difference between a cohort and a case-control study.
Cohort studies can be used to calculate relative risk (RR), incidence, and/or odds ratio (OR). Case-control studies can be used to calculate an OR.
Attributable risk?
The incidence rate (IR) of a disease in exposed – the IR of a disease in unexposed.
Relative risk?
The IR of a disease in a population exposed to a particular factor ÷ the IR of those not exposed.
Odds ratio?
The likelihood of a disease among individuals exposed to a risk factor compared to those who have not been exposed.
Number needed to treat?
1 ÷ (rate in untreated group – rate in treated group).
In which patients do you initiate colorectal cancer screening early?
Patients with IBD; those with familial adenomatous polyposis (FAP)/hereditary nonpolyposis colorectal cancer (HNPCC); and those who have first-degree relatives with adenomatous polyps (< 60 years of age) or colorectal cancer.
The most common cancer in men and the most common cause of death from cancer in men.
Prostate cancer is the most common cancer in men, but lung cancer causes more deaths.
The percentage of cases within one SD of the mean? Two SDs? Three SDs?
68%, 95.4%, 99.7%.
Birth rate?
Number of live births per 1000 population in one year.
Fertility rate?
Number of live births per 1000 females (15–44 years of
age) in one year.
Mortality rate?
Number of deaths per 1000 population in one year.
Neonatal mortality rate?
Number of deaths from birth to 28 days per 1000 live births in one year.
Postnatal mortality rate?
Number of deaths from 28 days to one year per 1000 live births in one year.
Infant mortality rate?
Number of deaths from birth to one year of age per 1000 live births (neonatal + postnatal mortality) in one year.
Fetal mortality rate?
Number of deaths from 20 weeks’ gestation to birth per 1000 total births in one year.
Perinatal mortality rate?
Number of deaths from 20 weeks’ gestation to one month of life per 1000 total births in one year.
Maternal mortality rate?
Number of deaths during pregnancy to 90 days postpartum per 100,000 live births in one year.
