9. Evidence Based Dentistry Flashcards
- Describe the main types of variables in oral health epidemiology
- Understand and interpret counts, rates and proportion
Categorical
2: Binary
If >2: Hierarchical → Ordinal
If not → Nominal
Counts- one particular point in time or over period of time
Proportions- %
Rates- frequency with which event occurs in a defined population in a specific period of time
In health epidemiology, gives idea of how quickly disease is dvping
Person year = the time people spend in period at risk of the disease (no. of ppl x no. of yrs)
—————————————–
Numerical
Decimal place- continuous
If not - discrete
Mean and median
- Interpret the concepts of prevalence and incidence and explain their relationship
For binary outcomes.
Incidence = New cases
How fast new cases arise in population
Counts, proportion, rate
Prevalence = Existing cases
Snapshot of cases in the population at a given point in time
Counts, proportion
…in a given period of time as a proportion of the total population in that same period of time
→ describe outcomes but not exposure-outcome relationships, do not effect measures taken.
- Explain why measures of dispersion are necessary when presenting data using measures of central tendency
Measures of central tendency can be skewed by outliers
Need to know extent of variability- range of data
What is a measure of effect
Measure of effect = quantity measuring effect of EXPOSURE on a health OUTCOME.
- Understand and interpret risk differences, risk ratios, prevalence ratios and odds ratios
- Explain the differences and applicability of absolute and relative measures of effect
Absolute measure: absolute amt an exposure adds/subtracts to health outcome- original units of the outcome are retained
Quantify excess risk or rate of disease in exposed grp vs unexposed
provides a measure of the actual number of people who might be affected and the potential public health impact of the exposure, providing focus on
How much impact would prevention have?
How many people would benefit?
a) Risk difference
Incidence of outcome among exposed / incidence of outcome among unexposed
—————————————
Relative measure: relative amt by which exposure increases/decreases risk/rate of health outcome.
Relative risk or rate of disease in exposed vs unexposed
Representing the strength of the association between an exposure and a disease by addressing questions such as:
How much more likely are exposed persons to develop the outcome than the unexposed?
a) Risk ratio/relative risk (incidence exposed/unexposed)
b) Odds ratio (odds of disease)
c) Prevalence ratio (prevalence outcome)
- Understand and interpret mean differences and beta coefficients
Linear regression draws a line for the average of values on scatter plot
Beta coefficient expresses the gradient of the line
Average change in outcome per one-unit increase in exposure
- Explain the differences between crude and adjusted measures of effect, and the reasons we need to consider confounding factors.
Observational studies are subjected to confounding
Occurs when all or part of the apparent association btw the exposure and the outcome is accounted for by other variables that affect the outcome n are assoc w exposure
Adjustment
To address confounding factors, researches adjust their measures of effect by stratification
Running regression models?
Creates ‘balanced grps’ on confounding factors = exchangeability
Interpretation of the tables:
The “crude” association between coffee consumption and periodontal disease was RR (relative risk) = 1.7
* When stratified by smoking, the association between coffee consumption and periodontal disease was RR = 1.0 among BOTH smokers and non-smokers
The crude association between coffee and periodontal disease was “confounded” by smoking.
Those individuals were all the same – or “exchangeable” - on smoking status - no
association between coffee consumption and periodontal disease among those with same smoking status → smoking could not ‘confound’ the association between coffee consumption and periodontal disease
