Epidemiology Flashcards
List the basic types of Epidemiologic studies
- Descriptive epidemiology (distribution, description)
2. Analytical epidemiology (hypothesized causal relationship, evaluating interventions, risk factors)
Recognize primary health outcomes (“the five D’s”)
Death Disease Discomfort Disability Dissatisfaction
** can add destitution (financial cost of illness)
prevalence
proportion of person in population who have the disease at a specific point
= number of existing cases of a disease / total population (at a given time)
describe sensitivity
Sensitivity (ability of a test to correctly identify those with the disease, aka true positive rate)
= True positives/(True positives + false negatives)
What are the major biases that complicate the interpretation of studies of screening results.
- VOLUNTEERISM: people who attend screening differ from those who don’t. The benefits of screening will exaggerate the benefits.
- LEAD TIME BIAIS: in evaluating the benefit of screening on reducing mortality, a bias in the comparison of survival tomes may result if lead time is not taken into account. In a RCT, we have to compare mortality, nor survival.
- LENGHT -TIME BIAIS: Screening will more likely detect those persons having a longer preclinical disease phase resulting in over representation of slowly progressing disease
however something is measured, we want the mesurement to be…
- Validity: does a measurement reflect the truth? (FREE OF BIAS)
- Reliability: precision, reproducibility
- Responsiveness: does it change as the phenomenon changes?
- Interpretable: Do people reading it know what it means?
Incidence
number of NEW events in a population of individuals AT RISK during interval
what are the 2 specific types of incidence mesures ?
- cumulative incidence
2. incidence rate (incidence density)
cumulative incidence
the proportion of individuals INITIALLY FREE of the disease of interest who develop the outcome during a defined time period
CI = # of new cases during a given period/total population at risk
Incidence rate
number of events divided by the amount of person-time observed, measures how quickly people are developing the disease
DI = # of new cases of a disease in a given period of time/total person-time of observation
what’s the link between prevalence, incidence and duration ?
prevalence = incidence x duration
- lower prevalence = high incidence, short duration (people who have the disease die fast)
- higher prevalence = low incidence/long duration (people live with the disease for a long time)
name 2 special types of incidence mesures and why they’re used
- mortality rate
- attack rate (a type of cumulative incidence , for accute epidemis)
why is prevalence mesure useful ?
- most useful for health care providers
- to assess the public health impact of a specific disease within a community
- to project medical care needs for affected individuals
why is incidence mesure useful ?
- assessing exposure disease relationship
- risk factors for disease
what are adjusted rates ?
used to be able to compare rates between two populations that are different in terms of some important characteristics (age structure, sex, etc.)
what would be the best Dx test ?
- valid
- reliable
- inexpensive
- easy to administer
- of minimal discomfort to the patient
what are the most useful test in clinical practice ?
those who produce the biggest changes from pretest to post-test probabilities
specificity
ability of the test to correctly identify those without the disease, aka true negative rate
= true negative / (true negative + false positive)
predictive value of a positive test (PPV)
probability that subjects with a positive screening test truly have the disease
= true positive / (true positive + false positive)
predictive value of a negative test (NPV)
probability that subjects with a negative screening test truly don’t have the disease
= true negative / (true negatives + false negatives)
do sensitivity and specificity change with prevalence ?
No, these remain fixed characteristics independant of what population the test is performed/disease prevalence
on what depend PPV and NPV ?
- sensitivity
- specificity
- prevalence of the disease
as prevalence increase, what PPV and NPV increase or decrease ?
PPV increase and NPV decrease
what are likelihood ratios ?
- a way to incorporate the sensitivity and specificity of a test in a single mesure
- LIKELIHOOD OF A GIVEN TEST RESULT IN A PERSON WITH THE DISEASE COMPARE TO THE LIKELIHOOD OF THIS SAME RESULT IN A PERSON WITHOUT THE DISEASE
LR+
Likelihood ratio of a positive test (LR+) if LR+ is 10, the people who have the disease are 10x more likely to test positive
= sensitivity / 1-specificity
LR-
Likelihood ratio of a negative test (LR-) if LR- = 0.1, people who have the disease are about 0,1 fois more likely to test negative than those who do. (so want it as close to 0 as possible)
= 1-sensitivity / specificity
what does high LR+ help to do
ruling in the disease = conclusive changes from pre to post-test probabilities
what does low LR- value help to do ?
RULING OUT the disease
a test with a high sensitivity but a low specificity produce
- no false negative
- many false positive
a test with a low sensitivity but a high specificity produce
- many false negative
- no false positive
how is, for any disease determined the optimal cut-off point ?
it has to be selected depending on the consequences of missing a few positive if the cut-off is set higher or of falsely classifying more negatives as positives if the cut-off point is set lower.
what’s the receiver operating curve ?
a plot of what happens to sensitivity and specificity as different cutoffs are used (for a particular test)
what’s the preclinical phase ?
the phase between disease onset and symptoms (possible screening)
what are the 2 steps of screening ?
- screening test
2. follow-up for definite diagnosis
the feasibility of screening is based on those requirements
- burden of the disease
- availability of effective treatment
- performance of test
what are the requirement for screening
- accuracy (high sensitivity and specificity)
- reliability (reproducible)
- safe and acceptable
- simple and inexpensive
best trial to compare a screening program ?
RCT - randomized control trials
what are the negative consequences of screening ?
- physical harm
- psychological harm
- false reassurance
- financial harm
what are the disease’s requirements for screening ?
- severe, common and perceived to be an important population health problem
- must be able to significantly alter outcomes such as death or severe disability
- significant lead time
what are the different types of data ?
- nominal (in categories)
- ordinal (some sense of order, but interval not specified)
- interval (inherent order with specified intervals)
what are the 3 ways of defining abnormality ?
- abnormal = UNUSUAL
- abnormal = ASSOCIATED WITH BAD OUTCOME
- abnormal = TREATABLE (target levels)
what’s the regression to the mean ?
when you retest someone with an abnormal result, particulary if extreme, the second test will usually come back closer to the mean, for that population
