Interpreting Data Regarding Correlations, Risk Factors, Prevention, and Therapy Flashcards
what is the correlation coefficient (r)
mathematical expression of strength and direction of correlation between variables
- range from -1 to +1
- +/-0.5 indicates a weak correlation
interpret:
a study of the amount of smoking and the incidence of lung cancer generates a correlation coefficient of r=+0.7
indicates a strong positive correlation
ie: increased amounts of smoking are strongly associated with an increased incidence of lung cancer
pearson product-moment correlation (r) is used for:
spearman rank-order correlation (rho) is used for:
continuous data
ordinal scale data
how is correlation represented
scattergrams
what is the correlation of determination
expresses the proportion of variance in a variable explained by another variable
r^2, expressed as a percentage
interpret:
correlating exercise levels with serum glucose levels generates r=-0.4, squaring this value results in a coefficient of determination of 0.16 (16%)
- indicates that 16% of the variance in serum glucose levels can be accounted for by the exercise level
- this correlation is weak and doesnt necessarily mean that increased exercise causes a decrease in serum glucose
what is regression analysis
linear regression
multiple regression
a mathematical technique used to predict the value of one variable based on another
–> only useful in highly correlated variables
linear: predicts one variable based on another
Multiple: predicts the value of a variable based on multiple covariates (ex rate of heart attack based on cholesterol, bp, weight, age) & adjusts data to eliminate effects of confounding vars
what are the ways to measure effects of risk factors?
odds ratio (OR) absolute risk (AR) attributable risk relative risk (RR)
what is the odds ratio
measures the degree of association of a risk factor with a disease or outcome
- ratio of the odds that a case was exposed to the odds that a control was exposed
- used when actual incidence of a disease is not measured, such as case-control
=odds that a case was exposed to risk factor/odds that control was exposed to risk factor
= (#obese diabetic/#nonobese diabetic)/(#non-diabetic obese/# non-diabetic non-obese)
=odds that a case was exposed to risk factor/odds that control was exposed to risk factor
odds ratio (OR)
what does odds ratio show
shows an association between risk factors and condition, but is a relatively weak indicator of causality
what is the absolute risk (AR)?
the risk over a given time period of contracting a disease
- synonymous with incidence
- done for control group and for experimental group
=number developing disease/total number
=number developing disease/total number
absolute risk (AR)
what is attributable risk?
the difference in risk between exposed and unexposed groups
-the amount of disease ocurrances that are attributable to the exposure
=AR (exposed) - AR (unexposed)
=AR (exposed) - AR (unexposed)
attributable risk
what is relative risk?
measures how many times exposure to a risk factor increases the risk of contracting a disease
-ratio of the absolute risk of disease among those exposed to absolute risk among those not exposed to the risk factor
- can only be calculated when incidence is measured (such as cohort study)
- -> shows not only association but is a more powerful indicator of causality
=AR of disease among exposed/AR of disease among non-exposed
=AR of disease among exposed/AR of disease among non-exposed
relative risk
measures how many times exposure to a risk factor increases the risk of contracting a disease
-ratio of the absolute risk of disease among those exposed to absolute risk among those not exposed to the risk factor
relative risk
the risk over a given time period of contracting a disease
- synonymous with incidence
- done for control group and for experimental group
absolute risk
the difference in risk between exposed and unexposed groups
-the amount of disease occurrences that are attributable to the exposure
attributable risk
Interpret:
a cohort study investigating obesity as a risk factor for diabetes determines a relative risk of 5
obese are 5X as likely to develop diabetes compared to non-diabetes
interpret
attributable risk of obesity to diabetes is 16%
16% of risk for diabetes comes from obesity
what are the ways to interpret data regarding prevention and therapy?
relative risk reduction
absolute risk reduction
number needed to treat
cost of intervention
what is relative risk reduction (RRR)
measures the decrease in relative risk due to an intervention
=1-relative risk
measures the decrease in relative risk due to an intervention is:
relative risk reduciton
=1- RR
relative risk reduction
interpret:
obese patients are treated with an aggressive diet/exercise program. compared to controls the treated patients are found to have a RR of 0.25. the RRR is thus 1-0.25=0.75. what does this mean?
treatment reduced the relative risk by 75%
what is absolute risk reduction (ARR)
decrease in absolute risk (decrease in incidence) of an outcome due to an intervention
=AR(Control) - AR (treatment)
=AR(Control) - AR (treatment)
ARR
decrease in absolute risk (decrease in incidence) of an outcome due to an intervention is:
ARR
interpret:
obese patients are treated with an aggressive diet/exercise program. controls are found to have an AR of diabetes of 0.4. The study group AR of diabetes is 0.1. The ARR is 0.4-0.1=0.3. what does this mean
the treatment reduced the absolute risk by 30%
what is the difference between ARR and attributable risk? (they are calculated the same way..)
- AR is generally used when assessing impact of risk factors
- ARR is generally used when accessing the impact of prevention/treatment
What is the difference between RRR and ARR?
- RRR is how many times exposure to a risk factor increases risk of contracting disease
- -> higher, more impressive number than ARR
- ARR is the decrease in absolute risk of an outcome due to an intervention
- -> smaller, less impressive than RRR
but ARR is more clinically relavant
why is ARR more clinically relevant than RRR
it incorporates the incidence of the disease/outcome being prevented
what is the number needed to treat (NNT)
calculation of the number of patients who would be required to receive treatment in order to prevent one negative outcome
=1/ARR
=1/ARR
NNT
calculation of the number of patients who would be required to receive treatment in order to prevent one negative outcome
NNT
interpret:
ARR for obese patients receiving aggressive diet and exercise was found to be 0.3. Calculate NNT. what does this mean?
NNT= 1/0.3 = 3.3
this means that 3.3 patients would need to receive treatment in order to prevent one case of diabetes
what is the cost of intervention
a calculation of the cost of preventing one negative outcome
=NNTx cost of 1 intervention
what is quality of life years (QALY)
measure of length of life weighed by a valuation of health related to quality of life
- quality of life modifier is calculated based on questions regarding ability to walk, self care, ability to perform usual activities, pain, anxiety, depression.
- generally this number ranges from 0-1, 1 is perfect health
