simplified formal definition Flashcards
Sensitivity
Proportion of true positives correctly identified as such by the test
False negative
True positives that are
incorrectly identified as negative
Specificity
Proportion of true negatives correctly identified as such by the test
(Specificity= 1- False positive)
False positive
True negatives that are
incorrectly identified as positive
Positive predictive value
Proportion of identified positives that are truly positive
Measure of process and
feasibility of running a screening programme
Confounder
A variable that is associated both with the exposure and the outcome and it is not on the causal pathway between exposure and outcome.
It gives an alternative
explanation for an association between exposure and outcome. It is nuisance and need to be controlled
Interaction
Effect modification
It is the situation where the association between exposure and outcome varies according to the level of a third factor
It is real effect that needs to be
detected and reported
Crude estimate
A measure of effect which has not been altered to take into account the effect of confounding factors
Adjusted estimate
A measure of effect which has been altered (by stratification or regression modelling) in order to take into account the effect of confounding factors
Regression analysis
Studies the association between two or more variables, where one is dependent on the other(s). This allows the dependent variable to be estimated given the value(s) of the other(s)
Direct (age) standardisation
– observed age-specific rates of death or disease from the study population are applied to standard population of known age structure, thereby calculating age-adjusted rate
Indirect (age) standardisation
– age-specific death rates of standard population are applied to the age structure of the study population and compare total number of observed deaths with the number expected
DSR
Directly standardised rate is total number of deaths (that would have been expected if the study population had the same age structure as the standard population) divided by total person-time of the standard population
SMR
Standardised mortality ratio is the ratio of observed to expected deaths obtained by indirect method of standardisation (usually expressed as a %)
p-value
It is the result of hypothesis testing. It is the probability of obtaining the observed or more extreme estimate if the null hypothesis (Ho) were true. Small p-value suggests that the null hypothesis is unlikely to be true.
Probability of making type I error i.e. wrongly rejecting Ho in favour of H1
95% Confidence interval
95% confident that the range of interval around the sample estimate contains the true population parameter
If the sampling is repeated thousands of times, the interval around the sample estimate will include the true population parameter 95% of the time
Power
Probability that an effect will be detected if it is truly there
Precision
Ability to measure magnitude of effect with minimal sampling error
Prevalence
Number of existing cases in a given time / total population at risk
Incidence risk
Number of new cases during a period of time / population at risk at the start of the period
Incidence rate
Number of new cases / total person-time at risk
Risk
It is the proportion of subjects who have experienced an outcome within a specified time period
Odds
Ratio of probability of
occurrence of outcome to nonoccurrence
Rate
Number of outcome events per unit person-time of follow up
Hazard
Rate defined at a precise point in time. Used in survival studies or Cox regression
In survival studies, it is
interpreted as the probability of dying at a particular point in time
Risk ratio
RR= Incidence risk in exposed / incidence risk in unexposed
Odds ratio
In case control study: OR = odds of exposure among cases / odds of exposure among controls
In cohort or intervention studies: OR = odds of outcome in exposed / odds of outcome in unexposed
Attributable fraction
Proportion of disease or outcome in the exposed individuals that can be attributed to the exposure
(measure of effect)
Attributable risk is used when there is causal association between exposure and outcome
Population attributable fraction
Proportion of disease or
outcome in the entire population that can be attributable to the exposure
(Measure of impact)
NNT
Number needed to treat
Mean
Sum of the observations divided by the number of observations
Mode
The most frequently occurring value in a set of data
Median
The value halfway through the dataset, below and above which lies an equal number of observation
Quintiles /
Deciles
Measure of spread
Outlier
An observation in a dataset that is substantially higher than lower than others in the dataset
Bias
Systematic error in the design or conduct of a study that results in conclusions that are different from truth
It could be:
Selection bias - error in selecting the study population such that comparison groups are not comparable
Information bias - error in measurement of exposure or outcome that results in systematic difference in the accuracy of the information collected between comparison groups. Includes reporter bias/ recall bias; interviewer bias / observer bias
Selection bias
- error in selecting the study population such that comparison groups are not comparable
Information bias
- error in measurement of exposure or outcome that results in systematic difference in the accuracy of the information collected between comparison groups. Includes reporter bias/ recall bias; interviewer bias / observer bias
Validity
The extent to which a test measures the true value of the variable of interest
Reliability
The extent to which a test will produce the same result if it is repeated at different periods or by different observers
Effectiveness
Effect of an intervention under operational conditions
Efficacy
Effect of an intervention under trial conditions
Efficiency
Effective outcome with use of minimal resources in terms of time and money
Systematic
review
Systematic assembly, critical appraisal and synthesis of all relevant studies on particular topic
Meta-analysis
A statistical method of pooling information from systematically selected studies and quantitative synthesis of summary result. It is used in order to give an
accurate estimate of effect and to overcome problems of reduced statistical power of small sample size studies
Blinding
Concealment of information about exposure or outcome in order to reduce bias
Randomisation
A procedure in which the probability of allocation to a particular group is predetermined and each individual has equal chance of being allocated to all groups. Randomisation ensures that intervention and control groups are similar with respect to known and unknown confounders, and it prevents bias in the allocation of participants to intervention or control groups
Randomisation
A procedure in which the probability of allocation to a particular group is predetermined and each individual has equal chance of being allocated to all groups. Randomisation ensures that intervention and control groups are similar with respect to known and unknown confounders, and it prevents bias in the allocation of participants to intervention or control groups
Intention to treat analysis
Used in intervention studies,
comparing intervention and
control groups as they were originally randomised, irrespective of whether some people in each group stopped taking the allocated treatment, changed to other treatment, or were lost to follow up