Final Exam Flashcards
Steps of evidence-based practice
Formulate question, search for evidence, critical appraisal, apply to patient, evaluate performance
where is research applied in the steps of evidence-based practice
critical appraisal, apply to patient
Simplified steps of evidence-based practice
Formulate, search, appraise, decide, reflect
In own words, definition of evidence-based practice
Appropriately integrating the best available evidence (obtained through systematic and thorough literature review), patient preference, and clinical expertise to make decisions about patient care.
the 5 A’s of evidence-based practice
Ask, acquire, appraise, apply, assess/adjust
what are the types of EBP questions?
Diagnosis, treatment, prognosis, epidemiology
MeSH stands for
medical subject headings
steps for case control study
Identify group with the disease (cases), identify healthy people who resemble the cases (controls), examine histories of cases and controls to identify possible causes, determine proportion of cases to controls who have the exposure
how many controls should there be per case in a case control study
1-4
what should be used to study a rare disease?
case control study
what size population do case control studies require
relatively small
what are case control studies best used for
identify possible causes of a rare disease
which study type uses odds ratio
case control study
what is the odds ratio
the likelihood that an effect will occur with an exposure over the likelihood that it will occur without an exposure
odds ratio >1 means
increased occurrence of an event (correlation)
odds ratio <1 means
decreased occurrence of an event (protective exposure)
what is a cause
A condition that precedes an event such that if the condition were different, the event would not have occurred
Bradford Hill’s elements of sufficient cause
Biological plausibility, temporal relationship, strength of association, experimental evidence, dose-response relationship, replication/consistency
types of cohort studies
retrospective, prospective
equipoise
the point at which a rational and informed person has no preference between two courses of action
first RCT
streptomycin for TB infection in UK 1949
types of data
nominal, ordinal, interval, ratio
nominal data
puts things into named categories without hierarchy
ordinal data
puts thing into named categories with hierarchy but no assumption of equal intervals
examples of nominal data
eye color, gender, marital status
examples of ordinal data
improved/same/worse
interval data
ordered but with the same distance between sets and an arbitrary zero
examples of interval data
Farenheit/Celsius, IQ
ratio data
ordered with same distance between sets and a meaningful zero
examples of ratio data
Kelvin, weight, height, length
Factors in evaluating health data
Timeliness of data, how representative is the data of the defined population, completeness of data, data collection issues, agenda of collecting organization, accessibility of data
Incidence and prevalence are measures of
disease frequency
Prevalence
amount of disease in a population at a given time
Prevalence is expressed in
percentage or proportion
point prevalence vs period prevalence
point prevalence is one point in time, period prevalence is over a defined amount of time
formula for prevalence
people with disease/people in population
Incidence
the rate of new disease development over a given period of time
what is implicit in most definitions of incidence
the population was initially free of disease
cumulative incidence aka
incidence proportion
cumulate incidence formula
number of new diseases over time/population without disease at baseline X 100
what is cumulative incidence used for
to monitor a specific population over time
cumulative incidence is like (EMS)
continuous infusion
formula for incidence rate
number of new diseases over time/person-time at risk X 100
what is a problem with calculating incidence rate?
difficult to calculate person-time at risk for a large population…they use census data and report in cases/100,000, but it’s not exact
what is the only formula with time in the denominator
incidence rate
incidence rate is like (EMS)
drip rate for bolus
what types of studies determine prevalence
cross-sectional survey, descriptive
what types of studies determine incidence?
cohort, RCTs
another way to think of prevalence
incidence times duration
when should a highly sensitive test be used?
A case where a false negative would be really bad: Infectious disease, serious can’t miss diagnoses…also at the beginning of a diagnosis process to narrow the possibilities
when should a highly specific test be used?
to confirm or rule in a suspected diagnosis, or when a false positive could harm the patient
predictive value of test aka
post-test probability
goal in ordering diagnostic tests re: post-test probability
to make the post-test probability significantly higher or lower than the pre-test probability
predictive values are influenced by
prevalence
positive predictive value formula
(true positives)/(total tested positive) x 100
negative predictive value formula
(true negatives)/(total tested negative) x 100
sensitivity formula
true positives/total with disease x 100
specificity formula
true negatives/total without disease x 100
what is positive predictive value
the probability that a person with a positive test result has the disease
what is negative predictive value
the probability that a person with a negative result does not have the disease
what affect on positive predictive value does a low prevalence have
it makes the ppv low and specificity has a higher impact on PPV
what affect on npv does prevalence have
not very much
PPV is highest for ___ diseases
common
ways to measure post-test probability
predictive values, likelihood ratios, diagnostic criteria
what is a likelihood ratio
the likelihood of a given test result in a person with the disease compared with the likelihood of the same result in a person without the disease
likelihood ratio formula
sick people with given test result/well people with given test result
how to calculate positive likelihood ratio
sick people who tested positive/well people who tested positive
how to calculate negative likelihood ratio
sick people who tested negative/well people who tested negative
how to calculate positive likelihood ratio in sensitivity
sensitivity/(1-specificity)
how to calculate negative likelihood ratio in specificity
(1-sensitivity)/specificity
what likelihood ratio value is helpful in ruling in a disease because it largely increases post-test probability
> 10
what likelihood ratio value is helpful in ruling out a disease because it largely decreases post-test probability
<0.1
benefits of using likelihood ratio
not dependent on prevalence, can be used for non-dichotomous tests, less exaggeration of benefits of test compared with sensitivity and specificity
what is key to establishing pre-test probability
good history and physical
at what pre-test probability are tests most useful
around 50%
in sequential testing, the post test probability becomes
the pretest probaility of the next test
what is the null hypothesis
the initial assumption that the results in the study are no different than the results that would have occurred by chance alone
what does it mean when the null hypothesis is rejected
the observed differences between the groups are not due to chance
what does it mean when the null hypothesis is not rejected
there is no way to tell whether the differences are from chance or not
what does the p value measure
the likelihood that the result is due to chance alone (the probability that the null hypothesis is true)
what is the range for p values
0-1
the smaller the p value
the less likely the result was due to chance
if p value is <0.05
the results are statistically significant, the null hypothesis is rejected, and the results are likely not due to chance
if p value is >0.05
the results are statistically significant, the null hypothesis is not rejected, and it is unknown whether the results are due to chance
what do confidence intervals measure
precision
what do overlapping confidence intervals mean
that no likely difference would occur when treatments are administered to larger population
what does it mean if CI crosses 0
no statistically significant difference between groups being compared
what does a wide CI mean
results were imprecise, uncertain quality of data
what does it mean if CI does not cross 0
statistically significant difference between groups being compared
what is number needed to treat
measurement of the impact of a therapy
what is the ideal number needed to treat
1
what causes nnt to become larger
cointerventions that alter outcome
what nnt is considered acceptable for a symptomatic condition
usually 5 or under
what is number needed to harm
measure of harm caused by intervention when compared to control group
the larger the number needed to harm
the lower the likelihood of adverse effects
