CIC to review: Stats and Epi Flashcards
Chart used to compare values across multiple groups
Bar
Types of data used for bar graph
1 Q
1 C
Chart used to examine distribution of quantitative variable by splitting it into multiple groups
Histogram
Data types for histogram
2Q (one displayed as groups)
Bars represent mutually exclusive groups
Chart used to show changes over time
Line chart
Types of data for line chart
1Q, 1C
Charts used to show how each category contributes to the whole
Pareto
Pie
Types of data for pareto chart
1 Q, 1 C
Types of data for a pie chart
1Q, q C
Chart used to look for correlation between variables
Scatterplot
Data types of scatterplot
2Q
Attributable proportion
RR-1 / RR
Measure of extent to which the distributions of possible results under the research hypothesis and null hypothesis do not overlap
Effect size
subset of sample statistics used to estimate a point characteristic of the population (eg it’s central tendency) rather than it’s dispersion (range, variance, etc)
Point estimate
way of assessing how much independent information is available from you data to estimate the parameter or determine the shape of the distribution
degrees of freedom
Using false positive- what happens if p <= alpha?
the risk of a false positive is acceptable, we reject the null hypothesis and conclude a statistically significant difference exists
Type 1 error
alpha- false positive
Type 2 error
beta- false negative
parametric tests used to determine whether two or more groups differ from each other based on the variability of values within each group versus the variability of values bewtween the groups
ANOVA
A nonparametric test used to evaluate how close the observed counts are for a single variable to those expected if the data fit a specific distribution
Chi square goodness of fit test
A nonparametric test used to determine how different the observed counts are from those expected if there is now association between the tested variables
Chi square test of association (AKA pearson’s chi square or chi-square test for independence)
Tests measuring the strength and direction of the relationship between two quantitative variables
Correlation
Parametric correlation
Pearson’s correlation
Nonparametric correlation
Spearmans
Models used to determine whether the linear relationship between 2+ variables is statistically signficiant, to assess the strength and direction of that relationship, and to predict values of the dependent variable for given values of the independent variables
Linear regression
Models used to predict probability of a particular outcome based on the values of one or more independent variables. The outcome possibilities must be dichotomous (ie survived/ died; infection/ no infection)
Logistic regression
nonparametric alternative to independent samples t-test, used to determine whether two groups differ based on ordinal ranks
Mann-Whitney U-Test AKA Mann-whitney wilcoxin test
Nonparametric test used to determine whether two related groups differ based on responses to a dichotomous variable. It is often applicable to pretest/ posttest studies with matched pairs
McNemar test
Parametric test used to determine whether the average values of two groups (or one group and a test value) are different. Versions of the test differ depending on how many groups are being examined and whether or not they are independent of each other
t-test
non-parametric alternative to the paired-samples t-test, used to compare the median value of a group to that of another group
Wilcoxon signed-rank test
What report shows facilities with need for improvement?
TAP- targeted assessment prevention
What is the CAD?
Cumulative attributable difference- # of infections to prevent to meet goals
Elements of effective surveillance program
- Select methodology
- assess and define pops to be studied
- choose events to monitor
- Determine time pd for observation
- ID surveillance criteria (case definitions)
- ID elements to be collected
- Determine methods for data collection and mangement
- Design interpretive surveillance report
- ID recipients of surveillance report
- Develop written surveillance plan
- surveillance program evalulation
What should be considered when selecting an event for surveillance?
- type of hc setting
- pop being studied
- Procedures performed/ services provided
- acuity of care
- risk assessment
- regulatory requirements
- Available resources
- PH needs
- performance improvement initiatives
- organization objectives
Note- usually high risk/ high volume events
Considerations when writing surveillance report
- Define event, population, setting, time period
- case definition criteria
- numbers
- methods
- purpose
- interpretation
- action/ recommendations
Examples of process events
Compliance with IP protocols
- Isolation
- Safe injection
- HH
- env cleaning
Methods to compare surveillance rates
z-test
SIR
Surveillance summary measure used to track performance across groups over time
Standardized infection ratio
Measure used to target prevention by identifying which locations have highest burden of excess infections
CAD- cumulative attributable difference
Risk adjusted measure used to track device utiliztation
SUR- standardized utilization ratio
Summary measure comparing antimicrobial use within and across facilities to guide stewardship effforts
SAAR- standardized antimicrobial administration ratio
Surveillance that uses real time data for early response
Syndromic surveillance
Surveillance that collects data from sample reporting sites
Sentinel surveillance
Ability to detect specified difference?
Power
What impacts the power in a statistical study?
- sample size
- Significance level (alpha)
- true value of parameter being tested (higher effect size, higher power)
Hill’s criteria
Strength of association
Specificity
Temporality
Biological gradient
plausibility
coherence
experiment
analogy
Incidence of disease is higher in those exposed (example lung cancer common in smokers)
Strength of association
Association observed in numerous studies
Consistency
1 factor, 1 disease
Specificity
Exposure to hypothesized causal factor precedes disease onset
Temporality