Measuring The Disease: Study Deisng And Statistical Analysis Flashcards
(74 cards)
1
Q
Descriptive epidemiology studies
A
Case reports Case series Incidence Cross sectional Ecological
2
Q
Analytic epidemiological studies
A
Experimental
Observational
3
Q
Experimental analytic epidemiological studies
A
Clinical trial
Community
4
Q
Observational analytical epidemiological studies
A
Cohort -prospective -retrospective Case control Other
5
Q
Descriptive epidemiology
A
- describes the amount and distribution of disease within a population, WITHOUT REGARD TO CAUSALITY
- identifies to whom, when and where the disease is occurring
- observational, not experimental
6
Q
Highest prevalence of colro coins deficiency in boys but ethnicity
A
Caucasian
7
Q
Analytic epidemiology
A
Concerned with causes and effects of disease within populations
- associations between exposures and outcomes
- asks WHY and HOW disease is occurring
8
Q
Diet and AMD: Melbourne collaborative study
A
Cohort study
- diet high in fruits, vegetables, chicken, and nuts and low in red meat seems to be associated with lower prevalence of advanced AMD
- no particular food pattern associated with early AMD
Looking at exposure and outcome
9
Q
Types of descriptive studies
A
Aggregate
Ecological studies
Individual
Casereport
Case series
Cross sectional study
10
Q
Descriptive study designs
A
(Individual person)
Case reports
Case series
Cross sectional
Ecological (aggregate)
11
Q
Case reports
A
- detailed description of a single case
- often a unique case
- cannot generalize, but can imitate studies
12
Q
Case series
A
- subjects of common characteristics of a disease
- no healthy comparison group
13
Q
Postprandial transient vision loss
A
Some people with carotid artery disease will have vision loss after eating due to steal syndrome. Blood being stolen from the eye and being shunted to the mesenteric system, causing the eye to be hypoperfused
Example of a case report
14
Q
Cross sectional study
A
- examines relationship between disease and other variables in a defined population at a specific time
- also called “prevalence study”
- A SNAPSHOT OF THE STUDY POPULATION
- advantage-can be done in a relatively short period of time with large populations
15
Q
Framingham study is a
A
Cohort study
16
Q
Example of cross sectional study
A
Looking at ophthalmic disorders in 40 stroke patients
Not analytic, does not look at cause and effect
17
Q
Ecological study
A
Units of analysis are POPULATIONS OR GROUPS as opposed to individuals
- AGGREGATE RISK is determined
- may be done when group, but not individual data is known
- ecologically fallacy-findings may not necessarily be applied to the individual. Want to avoid this.
18
Q
Examples of ecological study
A
Correlation between dietary fat intake and breast cancer by country
Do not know whether individuals with high fat intake had breast cancer
19
Q
Analytic study types
A
Experimental studies
-Randomized control (intervention) trials
Person doing study intervenes
Observational studies -Cohort -Case control -Cross sectional No intervention
20
Q
Experimental studies
A
Investigation manipulates one or more risk factors and analyses its effects
- can control external factors
- can provide strong evidence
- more expensive and difficult
- E.g. randomized controlled trial
21
Q
Observational studies
A
- people are observed to see whether there is a relationship between a risk factor and health status
- most common design used in epidemiology (cohort, case control studies)
- NO INTERVENTION GIVEN
- glaucoma being more prevalent in AA, AMD being more prevalent in whites, populations having various risk factors, etc.
22
Q
Cohort
A
Group of people with given characteristics followed over time (longitudinal study)
23
Q
Cohort: prospective study
A
- NONE of the individuals have the disease in the beginning
- follow into future to observe for presence or absence of disease
- compared risk factors between those who did and did not develop the disease
- INCIDENCE RATE is looked at
Look at a study population and follow them over time. Looks at those who were exposed to a certain a risk factor and those who were not. Seeing who gets the disease and who does not with respect to exposure of risk factor
24
Q
Example of cohort
A
Framingham
- 20 year cardiovascular mortality rates
- looking at people who died from cardiovascular disease in a 20 year period.
- no one had cardiovascular disease in the beginning
- the later the cohort, the less mortality rate there was
- seem to be less death from cardiovascular disease and mortality rate lower in females further on
- doesnt talk about cause of it
25
Retrospective cohort study
- identify individuals WITH AND WITHOUT DISEASE IN THE PRESENT
- go back to a time when ALL OF THESE WERE FREE OF DISEASE
- compare the groups with regard to risk factors
- often done by chart review
- incidence measured
26
Example of retrospective cohort study
Statin use and cataract surgery
- record review, 50,000 patients via national data base initially WITHOUT CATARACT OR STATIN USE
- statin therapy modestly assocaited with increase risk fo cataract surgery
27
Advantage of cohort study
-good way to evaluate relationship between development of disease risk factor
28
Disadvantages of cohort study
Not useful for diseases that take a long time to develop
Expensive
29
Case control studies
- those with the disease (CASES) and those without the disease (CONTROLS) are compared with regard to suspected RISK FACTORS for the disease
- medical records, historical data, interviews
- basically retrospective but not cohort
- usually done with small groups of people
Not following them over time to see if they develop a disease
-seeing if they have risk factors
30
Example of case control study
Acanthamoeba keratitis in SCL wearers
-patients more likely than controls to use homemade instead of commercially prepared saline and wear they lenses while swimming
31
When do you use case control over cohort
For diseases that are more rare and you don’t have a big population to follow. Cohort used for more common diseases
Congential cataract following German measles in the mother
32
Advantages of case control studies
Inexpensive
No need to follow over time
Looking at records
Useful in studying rarely occurring diseases
33
Disadvantage of case control studies
Small number of people
Difficult to generalize
No new cases, so incidence is not measured
34
Probability of an adverse event taking place in a population within a specified time
Absolute risk
35
What is absolute risk a measure os
Incidence
36
How is risk quantified in epidemiological studies
Absolute risk
37
A new CL wearer asked you, “what is my risk of getting a corneal ulcer”
Even if he is not a CL wearer, he may get an ulcer
- the concept of absolute risk does not distinguish between the risk fo ulcers for CL wearers and the risk of ulcers for non CL wearers
- what he is really asking is what is the EXCESS RISK?
38
How do we measure the excess risk
- also known as the RELATIVE RISK or the RISK RATIO
- risk (incidence) in people exposed/risk (incidence) in people not exposed
- lets follow CL wearers and non CL wearers for 10 years and see who develops corneal ulcers. None have ulcers to start
39
Relative risk of smokers to non smokers of getting AMD
3.29x in smokers compared to non smokers according to Beaver Dam
40
Smoking and risk of AMD in men
Current smokers >20 cigarettes per day compared with never smokers=RR 2.46
Past smokers > 20 cigs a day RR 1.30
Current smokes < 20 cigs RR 1.26. Not stat sig
41
Relative risk in a cohort study: risk of CHD by triglyceride level
The higher the triglyceride levels are, the greater the RR for CHD
42
Risks ratio is used in what studies
Cohort, not case-control studies
- case-control subjects are not followed over time to see whether or not disease occurs so we cannot directly measure the incidence rate
- but we can measure the frequency of exposure between cases and controls
- this ESTIMATES the incidence or relative risk
43
A comparison of the frequency of exposure among cases AMD controls
Odds ratio
| -smoking, CL use, exercise, a medication, meditation etc
44
Odds ration is used in
Case control
Cross section
Cohort
45
How else can odds ratio be expressed
Exposure or disease odds ratio
46
Exposure odds ratio
odds in favor of exposure among cases/odds in favor of exposure among controls
- it asks “ is a patient with a disease more likely to have a risk factor than a patient without the disease have the risk factor?
- example-the odds that a patient with an ulcer wears CLs/odds that a patient without an ulcer wears CLs
- the odds are 5x greater that a pateitn with an ulcer wears CLs than a pateitn without an ulcer wears contacts
- used in cross sectional or case contro lstudies
47
Disease odds ratio
- the odds in favor of disease among exposed/the odds in favor of the disease among unexposed
- it asks “is a patient with a risk factor more likely to have a disease than a patient without the risk factor have the disease
- the odds that a patient who wears CLs has an ulcer/the odds tha ta patient who does not wear CLs has an ulcer
- the odds are 5x greater that a pateitn who wears CLs has an ulcer than for a patient who does not wear CLs to have an ulcer
- used in cohort and cross sectional studies
48
Odds ratio and risk ratio similarity
Similar to risk ratio
>1.00-exposure results in greater risk
<1.00-exposure results in less risk
=1.00 exposure results in no risk
Exposure~risk factor
49
Example of odds ratio
Refractive error and strab
Hyperopia prevalence of esotropia with ORs increasing from 6.4 for 2D to <3D of hyperopia, to 122.0 for > 5D. Exotropia was associated with prematurity, maternal smoking during pregnancy, family Hx of strabismus, female sex, astigmatism
Conclusion: because refractive error is correctable, these risk associations should be considered when developing guidelines for the screening and management of refractive error in infants and young children
50
Blue mountain study about AV nicking
Odds are 4x greater that a person with a vein occlusion has AV nicking than a person without a vein occlusion
This is analogous to saying “the relative risk of vein occlusion is 4x greater in patients with AV nicking
51
Confidence interval
- CI is used to address the repeatability of the result, or how much it would vary from one study to another
- another way of saying this is the precision of study
- indicates the chance that the interval includes the true value
- so the true odds ratio for diastolic is between 1.082-2.06 with a 95% confidence
- CI depends on the sample size and the variability of the data
- the more narrow the CI the more precise the result
52
Study results/chance
- do your results reflect a true effect or ar they due to chance
- chance=a random error, inherent in all studies, can be minimized, never elimainted
53
Null hypothesis
The idea that there is no real difference, and that any statistical differences are due to chance alone
54
P value
The proababiltiy that the effect is due to chance alone
- in order to be soldiered statistically significant (reject the null hypothesis) P values typically start at P < 0.05, though this is arbitrary
- P>0.05 means there is a greater than 5% probability that the results are due to chance: fail to reject the Null hypothesis
55
RCT
A type of experimental study used to evaluate a preventative or therapeutic procedure or intervention
- evaluates the efficacy, as well as potential harm of intervention
- generally most scientifically rigorous method
Drug studies, VT studies, etc
56
Allocation of a RTC
Subjects randomly allocated into groups
- treatment-receiving intervention
- control-not receiving the intervention
57
Sampling in a RTC
- simple random sample-each person has an equal chance of being selected. E.g assigning numbers
- stratified random sample-dividing the population into subgroups according to impartiality characteristics; age, sex, race, location, economic status; the selecting random sample out of each subgroup
58
Bias in RTC
A systematic error in the design, conduct, or analysis of a study that influence the association fo an exposure on the outcome
59
Selection bias
Arriving from the selection of individuals
| -E.g. selecting subjects for a study solely through internet ads, newspaper ads, phone calls, in person-street locations
60
Informational bias
- accuracy of information is not the same for each group
- can results from errors in measurement, data collection, interviewing, classification among subjects (Cases vs controls)
Every person has to gather information in the exact same manner as another one
61
Confounding bias
Bias due to association of othe rfactors that influence the outcome
-example: is coffee drinking a risk factor for heart disease?
Smoking and coffee drinkers are more likely to smoke, smoking increases the risk of heart disease. In this case, smoking beceoms confounder.
62
Attempts to avoid bias
Masking
-simple blind: subjects do not know which group they are assigned to
-double blind: neither subjects nor examiners know which group subject is assigned to
AREDs was randomized, placebo controlled, and double blinded
63
Internal validity
The extend to which results reflect the relationship between the espxoure and outcome
64
External validity
The generalizability of the study to similar populations
65
Good studies have ____ validity
High
66
DREAM study
Does the omega 3 FA supplemation help the symptoms of dry eye
-multi center, randomally assinged, double blind, placebo controlled clinical trial
Omega 3 FA did not have significantly better outcomes than those who were assinged to receive placebo.
67
OHTN study (OHTS): does lowering IOP in persons with high IOP but without glaucoma delay or prevent them from developing glaucoma in the future?
- risk of developing glaucoma is halved by lowering IOP (RR=4.4/9.5=.46)
- but we can also measure the additional risk by subtracting the two risks. This is called the risk differnece (RD) aka “abolsute risk reduction or attributable risk=9.5-4.4=5.1%
- the number needed to treat is a way of specifying the benefits of treatment for the population studied based on the RD
- NNT=1/risk difference=1/.051~20
You would have to treat 20 patients with OHTN to prevent 1 case of glaucoma developing in 5 years
Puts results into perspective
68
Clinical trials for new drugs
Phases I-IV
69
Phase I clinical trial
Small group of people for 1st time
| -safety, stage range, and side effects
70
Phase II of clinical trials
Larger groups, efficacy and safety
71
Phase III clinical trials for drugs
Efficacy, side effects, COMPARISON WITH COMMONLY USED TREATMETNS
72
Phase IV clinical trial for new drug
Effect in various populations and side effects associated with long term use
73
Rhopressa
Did not meet its primary endpoint of being non inferior to timolo in patients with baseline IOP 20-26 in phase III
74
Phase III or Rhopressa
- changed the primary endpoint-now patients with IOP 20-25
- September 2015 results successful
- Rhopressa approved in December 2017 Nd launched April 2018
- roclatan: Rhopressa combined with latanaprost, new drug application acceptance this year
