Unit 1 Flashcards
Epidemiology
The study of distribution of disease and determinants of disease frequency in populations
Goal of Epidemiology
Is to control health problems and improve health at the population level
Operationally
Counting the causes of morbidity and mortality
Determining variables associated with causes of morbidity and mortality
In operationally why identify factors that are causes?
They are potentially modifiable.
-Guiding and evaluation interventions to improve public health
Basic Assumptions of Epidemiology
1) Death disease and disability do not occur at random
2) There are causal factors that can be identified through the systematic investigations of human populations
3) Identifying these causal factors can lead to preventative intervention
Clinical and Research Concerns: Exposure
Good or bad: Chemical, biological, physical, psychological, educational etc.
Clinical and Research Concerns: Outcome
Good or bad: Disease, cure, improved attitude, longer life, better QOL
We generally know either the exposure or the outcome and_____________
Want to measure the other
Endemic
The usual occurrence of a disease in a given population
Epidemic
A meaningful increase in the occurrence of a disease in a given population
Pandemic
Spread of a disease across a large region or worldwide
Epidemiological Reasoning
1) Suscpicion of an E–> D relationship
2) Hypothesis formation
3) Test E–>D hypothesis
4 Rule out alternative explanations: Chance, bias, confounding
Epidemiology is fundamentally concerned with________
Populations
Measuring distributions of disease in populations and the factors associated with those distributions
Association
An identifiable relation between an exposure and a disease
Methodological question
How do we look for a cause
Ontological question
What is a cause
Ethical Question
How do we decide if there is enough evidence to act on a cause
What is a cause
A cause of a disease is an event condition or characteristic the preceded the disease and without chick the disease WOULD NOT have occurred at all or would not have occurred until some later time.
Criteria for assessing Causality
1) Strength of association
2) Dose-response relationship
3) Temporal Sequence
4) Biological credibility
5) Consistency of findings across studies
Strength of Association
Is there a strong E-D relationship
BUT doesn’t imply that a weak association cant be judged as cause and effect
Dose-Response Relationship
Does risk increase with increase exposure?
Temporal Sequence
E–> D
Does the exposure precede the disease
Biological Credibility
Is there a known biological basis for the relationship
BUT DEPENDS ON CURRENT STATE OF KNOWLEDGE
Consistency of Findings
Do multiple studies report similar findings
Studies can differ by:
Investigator
Methodology
Study population
Epidemiologic Approach
Identify DISEASE Identify EXPOSURE Statistical associations between E/D Hold constant factors that may be mixed up in this measure of association Inference causal association Recommend intervention
Risk Factor
A factor whic if present increases the probability of disease occurrence
- The exposure must precede disease onset
- Must be associated with an increase disease frequency
- Absence of error and bias
Quantification in Epidemiology
Measuring diesel occurrence is fundamentel.
4 Types of scales
Nominal Scale
Ordinal Scale
Interval
Ratio
Nominal Scale
Uses names
Ordinal Scale
Follows an order based on severity
Interval Scale
Follows a mathematical order but has NO TRUE ZERO
Ratio Scale
Follows a mathematical order HAS a defined true zero
Quantification in Epidemiology
Depending on the time element we can also quantify cases as prevalent or incident by measuring:
- Prevalence proportion
- Incidence rate
Prevalence Proportion
=#Cases/#person in population ) at a specific time
Specific time can be a point or period of time
Ex. 45 D1 students /105 have at least one active carie
Incidence Rate
=# of new cases of disease/population at risk) over a time period
Ex.
7 new cases of periodontal disease per 105 D1 in 2010
_____Is not a rate but ______ is and is not meaningful without a time unit
Prevalence
Incidence
Issues with Incidence and prevalence
How do we know someone is a case
How do we count population at risk
What specific time period
Evidence Based Dentistry
EBD is an approach to oral heath care that requires INTEGRATION clinically relevant SCIENTIFIC EVIDENCE, relating to the patients oral and medical condition and history with the DENTISTS CLINICAL EXPERTISE and the PATIENTS TREATMENT NEEDS AND PREFERENCES
EBD is based on 3 components
1) Best available scientific evidence
2) A dentists clincal skill and judgement
3) Patients needs and preferences
What approach does EBD take
Patient centered!
It is an approach to practice and to making clinical decision. Just one component used to arrive at the best treatment. Provides personalized dental care based on the most current scientific knowledge.
EBD IS:
Systematic approach to practicing good dentistry:
Provide right care
To the right patients
At the right time
EBD is NOT
- Cook Book Dentistry
- A standard of care
- Only about randomized trials (EBD process is not a rigid methodolical evaltion of scientific evidence)
- Impossible to practice
EBD: Outcome
Better quality of patient care and outcomes
-Evaluation of science underlying clinical care
-Quality of decision making
-Reduciton in variations in clinical practice
Improvements in research and dissemination of results
EBD can facilitate_______
BUT does not guarantee, making better declines in the provision of dental care
EBD Steps *5
1 Formulate a relevant question
2 Find the best available evidence
3 Review the evidence for its validity and applicability
4 Integrate the best research evidence with your clinical expertise and patients needs/desires
5 Evaluate your efforts
PICO
Anatomy of well formulated question:
P-population
I-intervention
C-comparison
O-Outcome
Quality of information (from best to worst 6)
1 RCT 2 We'll controlled cohort studies, case control and nonrandomize clinical trials 3 Cross sectional studies 4 Descriptive surveys 5 Case Reports 6 Personal opinion
Primary Source
Must be the first disclosure containing sufficient information to enable peers to assess obseravations repeat experiments and to evaluate intellectual processes
Secondary Source
Includes most books, review, articles, and indexes to the literature and usually summarize reviews or organize information
Conducting a Search
1 Defining the question
2 choosing a resource
3 Keyword vs classification systems
4 Search techniques
Defining the question
The most difficult step
PICO
Patient, intervention, comparison, outcomes
Boolean Operators
AND, OR, NOT
Truncation
Allow you to place a symbol at the end of a word which tells the engine to retrieve the word stem with different endings (usually an *)
Phrases
Use quotation marks this limits the results to the exact phrase
Parentheses
Use to appropriately group the terms and operators to control the order of the search
Explode
Instructs the search engine to retrieve the information with a broad subject heading that are broken down into narrower subject headings
Internal Validity
The degree to which the results of a study are likely to approximate the truth
External Validity
The extent to which the effects observed are applicable to a broader population
Inference can only correctly be made to the population from which the sample was drawn
We often leap to inference far beyond the targeted population
Observational Studeis
They observe the outcomes without intervening to affect them
Experimental Studies
The researcher manipulates the exposure(usually a drug or treatment) to compare it to the standard of care
Observational Studies (ex)
Cohort studies
Case Control Studies
Cross Sectional Studies
Cohort Studies
Subjects are selected based on their exposure status
2 Types of cohort studies
Prospective Cohort-compares disease prevalence in the exposed and unexposed
Retrospective cohort- They begin with the exposure of interest and probe back for exposure information
Advantages of Cohort Studies
Maintains temporal sequence (assesses exposure before outcome)
Good for assessing rare EXPOSURES and rapidly fatal disease
Can study multiple diseases/outcomes from a given exposure
Can calculate INCIDENCE among exposed and unexposed
Minimizes error in ascertainment of exposure(least prospective)
Provides complete description of experience after exposure including rate of progression and natural history of disease
Disadvantages of Cohort Studies
1) Expensive
2) Inefficient for rare diseases
3) Long follow up
4) Secular trends may influence behavior and study characteristics over time
5) Diagnostic trends
Cohort: Selection of “exposed” and “unexposed” depends on:
Research question
Ability to collect exposure and disease information
How common/rare the exposure
Unexposed
Similar to exposed other than exposure factor
Appropriate selection and investigation methods to avoid bias and confounding
Case-Control Studies
Subjects are selected based on their disease status
Case studies should theoretically mimic ________ studies
Cohort studies
In case control studies__________ are compared to________
Diseased people (cases) Non-diseased people (controls)
Cases and controls should be:
Different only on their past exposure
Case control Studies
Can demonstrate risk INDICATORS and not risk FACTORS due to the retrospective nature of the study design (temporality cannot be assessed)
Case Control Studies: Exposure
Cases and controls must have had an EQUAL CHANCE OF BEING EXPOSED
The exposure has to be assessed retrospectively and proportions of cases and controls who are exposed are unknown at the beginning of the study
Case Control Advantages
1) Efficient for rare diseases**
2) Relatively efficient in terms of time and money
3) Can study diseases with long latency period
4) Allow for the evaluation of multiple exposures that may increase risk for a specific disease
Case Control Studies: Disadvantages
1) cannot DIRECTLY compute incidence of diseases in exposed and non-exposed persons
2) Temporal relationship cannot be established wit certainty
3) Prone to errors in selection of cases/controls and in errors pertaining o the collection of information
4) Not optimal for rare exposures***
Case Control: Selection of Cases and Controls
?
Cases: Cases (disease) definition Diagnostic criteria Hospital based or population based Incident or prevalence
Controls:
Would be cases if had the disease
Potential for bias and confounding
Cross Sectional Studies
Selection of subjects based on neither exposure or disease status
- Most basic study design
- Point in time or snapshot information
- Subject selected without regards to exposure or disease status
- Does not need explained etiologic objectives
Cross Sectional Studies: Advantages
Sampling and analytic methods provide for statistically valid inference to populations
Exposure and disease are assessed at the individual level
Cross-Sectional Studies: Disadvantages
Temporality cannot be assessed
Experimental Studies (2 types)
Randomized Clinical trials
Community Intervention Trials
Randomized Clinical Trial
Principles of all experimental studies follow those from clinical trials
Have a long history in clinical medicine
Are sub-types of Cohort studies in which exposure is randomly assigned by the investigator
Randomization
The process by which each participants treatment is determined by some random mechanism
The PRIMARY PURPOSE of randomization is the minimizing of confounding
Why Randomize?
- To create groups (experimental and control) that are not determined by any other factor other than chance
- Minimize confounding (known and unknown)
RCT: Blinding
The investigator and or participant do not know what arm the participant is in
Single Blinded
The participant does not know but the investigators does know treatment assignment
Ex.
Toothpaste trial where participant does not know if she is using the experimental or the control paste
Double blinded
Where neither the participant nor investigator know treatment assignment
EX. Drug trials, where participants simply take a blue or red pill and no one knows which pill is which; data are coded and code only broken outside by monitors or at the end of the study
Purpose of Blinding
Removes bias or systematic error
Information Bias
Drawing different conclusions depending on their knowledge of which study arm particular participant is in
Selection Bias
Study recruiters can be eager to recruit “sick persons” into experimental arm
Key Elements of RCT
- Selection of study population
- Allocation of treatment/intervention
- Study conduct and compliance
- Follw up and establishing outcomes
Considerations in Experimental studies (3)
1) stopping rules
2) Sample Size
3) Analysis and Interpretation
Systematic Review
Systematic complete summary of the literature
Meta-Analysis
Combined analysis of data from different studies following strict guidelines
Scales of Measurement Data Types
Nominal Ordinal Continuous -Interval -Ration
Nominal
A measurement scale based on the classification of an observation according to the group to which it belongs
Ex. Gender, political party, marital status
Ordinal
A measurement scale based on the classification of an observation according to its relationship to other observations
Ex. Poor-fair-good rating scale
“Uneven Intervals”
Just because numbers are used doesn’t mean that they represent “true numbers’ that can be added and subtracted.
Numbers can represents nominal or ordinal scale values
Interval
A measurement scale characterized by equal units of measurement
- The distance between any 2 numbers is known size
- Zero point is arbitrary
Ex. Fahrenheit/Centigrade temperature scales
Ratio
A measurement scale characterize by equal units of measurement and a true zero point at its origin
Ex Mass, time
Populations Vs. Samples
Populations (Greek symbols)
Constants, not variables
Samples
Roman characters X, S (or SD)
Variables
Measures of Location or Central Tendency
Mode
Median
Mean
Mode
The value of the most frequent measurement
Useful with the nominal scale but may be used with any
There can be more than one mode
Median
The value of measurement that falls in the middle when the measurement are arranged in order of magnitude.
-Point at which 50% of the measurement fall
-Not useful with ordinal scale, but may be used with higher order scales
Mean
The arithmetical average: The sum of the measurements divided by the total number of measurements
Most useful with interval or ratio measurement
Measures of Variability or Dispersion
Ranges Interquartile Range Variance Standard Deviation Coefficient of Variation Standard error of the meant
Range
Range=Xmax-Xmin
Interquartile Range
IR=X25th -X75th
Variance
The average of the square of the deviations of the measurements about their mean
Considered in terms of distance of each measurement from the mean
“Degrees of freedom”
S^2
How to Calculate Variance
1) Find the mean
2) Subtract this sample mean from each of the measurements and squaring the results to eliminate negative numbers.
- The sum of these=sum of squares (ss)
3) Dividing ss by the number in the sample minus one gives the sample variance
Substracting one from the sample total=UNBIASED estimate of POPULATION variance
Standard Deviation
“Positive square root of the variance”
-The standard deviation is calculated by taking the square root of variance
Subtracting one from N gives us an unbiased estimator or the population standard deviation
Coefficient of Variation (CV)
Measures the percentage of spread
Unitless
Allows for comparisons
CV=100*SD/X
Central Limit Theorem
The sampling distribution of the mean of any independent random variable will be normal if the sample size is large enough
The More closely the sampling distribution needs to resemble a normal distribution the more sample points required
The more closely the original populations resembles a normal distribution the fewer sample points will be required
Standard Error of the Mean (SE)
Took several samples from a population, calculated the mean of each sample, then calculated SD=SE
SE may be calculated from a single sample by divine the SD by the square root of N
_______Is an estimate of the variability of individual measurements within a sample
SD
SD is used to measure variability of individual/subjects around a sample mean
_______ is an estimate of the variability of sample means about the population mean
SE
SE is used to assess how accurately a sample mean reflects a population mean
Interpretation of Mean (anesthesia study)
Best scientific guess of how long it takes to achieve anesthesia
Interpretation: SD
How much variantion there is in onset time
Interpretation: SE
If experiment repeated the true population mean would be within 1 SE
3 Ds of Epidemiology
Distribution
Determinants
Descriptive
Determinants are influenced by
Heredity Biology Physical environment Social environment Lifestyle
Descriptive (epidemiology)
Prevalence
Severity
Age adjusted distribution in population
Analytic
Try to answer a specific question
Primary Data
Mail survey
Epidemiological screening
Telephone interview
Secondary Data
Medicaid
Vital statistics
Cancer Registry
Chemical Acid Theory
17th and 18th Centuries. Decay arises from acids formed in the oral cavity. Assumed acids were inorganic
Parasitic (septic) Theory
Microorganism inifiltrae the enamel, leading to decomposition. Recognition that enamel is organic
Non-Specific Plaque Hypothesis
Total plaque microflora
Specific plaque hypothesis
Only a few species involved
Ecological Plaque Hypothesis
Shift in hemeostatic balance
Extended ecological Plaque hypothesis
Non-pathogenic bacteria can adapt to produce acid
Dental Caries
An infection communicable disease resulting in destruction of tooth structure by acid forming bacteria found in dental plaque, an intra oral biofilm in the presence of sugar.
Epidemiological triangle
Host
Agent
Environment
Agents of Caries
Streptococcus mutants (most cariogenic for enamel)
streptococcus sobnrinus
Vile lonely
Mutans Streptococci (ms)
Greater ms counts greater caries prevalence Caries conducive because -Ability to adhere to tooth -Produce copious amounts of acid -Survive at low pH
Hope would House Study
Sucrose restricted diet a outing 81 children age 4-9
At the start 78% cavity free
53% continued caries free at age 13
This was significantly higher than caries free 13 yers old within the general residential population-only .4%
Dental Caires Risk Factors
Age Gender Race and Ethnicity Socioeconomic status Geography
Deterrence
Saliva Plaque removal -OH Dietary Habits Fluoride Therapy Sealants Caries Vaccine Antibiotics Other
How To measure Dental caries
Counts
Proportions
Rates
Indices
Prevalence
The number of people in a population who have a given disease at ag vein point in time. Prevalence measures the frequency of all current cases of disease (old and new)
Incidence
A measure of the number of lesions/period of time
Limited in that they only measure the numbers of new initial lesions per unit of time
DMFT (DMFS)
Decayed
Missing
Filled
Teeth/Surface
Describe the amount of prevalence of dental caries in an individual
Maximum value for DMF(S)
128
DMFS value of ______ or higher can be considered to indicate sever disease in children up to age 17
7
Limitations of prevalence
Provides pas history only
Does not provide rate of lesion development
Does not indicated if caries is active or inactive
Does not provide the frequency of occurrence of new lesions
D+M+F
Prevalence
D/DMF
Untreated caries
F/DMF
Treated caries
M/DMF
Tooth fatality
DEFT or DFT
Decayed
Extraction
Filled
Index is useful up to the age of 6
Significant Caries Index
Bring attention to the individual with the highest caries values in each population under investigation
SiC index
Individuals are sorted according to their DMFT values
One third of the population with the highest caries scores is selected
The mean DMFT for this subgroup is calculated. This value is the SiC index
Overarching Goals of Healthy People 2010
Increase quality and years of healthy life
Eliminates health disparities
28 Focus areas
467 specific objectives
Practical Significance of the epidemiology of Dental Caries
Planning, funding, and delivery of services
-Example water fluoridation, clinics, Medicaid
Training: numbers and types of professionals
Why are Statistics important
It allows us to understand information and make clinical decisions based on DATA from:
Scientific journals
Clinical study reports
Product manufacturers
Presentations at dental conferences
Mean
Average of the date; sensitive to extreme values
Median
Middle point of the data; less sensitive to extreme values
Mode
Most frequently occurring value in the data
SD
Measure how much the individual data points vary around the mean
Frequency
Count of a given outcome or in each category
Percentage
Count of a given outcome per hundred showing proportion of each category out fo the total
Correlation
Is there a linear relationship between an independent variable X and a dependent variable Y
Value of correlation coefficient can lie between
-1 and +1
+ Value
As x increases, Y increases
- Value
As x increases y decreases
The closer r is to +1 or -1__________ the relationship
Stronger
Square of correlation
Is the fraction of variation in Y and explained by X
The higher the r^2 the better the fit of the regression line
Hypothesis Testing
A hypothesis is an explanation for certain observations. We use hypothesis testing to tell if what we observe in the population is consistent with the hypothesis
Null Hypothesis H0
Usually staters that there is no difference between 2 groups being compared or no effect of a product or intervention
Alternative Hypothesis Ha
States that there is a difference between 2 groups being compared or an effect of a product or intervention
Type 1 Error
Rejecting the null hypothesis that is actually true in the population
Alpha statistical significance
Type 1 error
Alpha is commonly set to .05 or 5% of incorrectly rejecting the null hypothesis when it is actually true
Type 2 Error
Failing to reject (accept) the null of hypothesis that is actually false in the population.
Probability of a type 2 error is beta
P-Value
The probability assuming that the null hypothesis is true of seeing an effect as extreme or more extreme than that in the study by chance
Reject null hypothesis is P-value _____ Alpha
Less than
Fail to reject (accept ) null hypothesis if P value is _____ alpha
Greater
Confidence Intervals
Is a range of values about a sample statistic that we are confident that the true population parameter lies
T-Test
Statistical test that can be used to determine whether the mean value of a continuous outcome variable differs significantly between 2 independent groups
Alternative test may be
Directional or non directional
One Sample T Test
Can be used when the outcome variable of interest is only being examined in one group
Matched pair T test
Can be used when subjects are matched in paired and their outcomes are compared within each matched pair ( including where observations are teen on the same subjects before and after a given intervention(
Chi-Squared Test
When examining categorical data, the chi square test can be sued to compare the proportion of subjects in each of two groups who have a dichotomous outcome
ANOVA
Analysis of Variance is a sticaltcal method that allows for comparison of several population means
F statistic
Used by ANVOA, rejects null hypothesis that the population means of all groups are equal if Pvalue of F statisistic is greater than alpha (.05)
The Fundamental Issue
Quantifying our confidence on how well the findings reflect the truth. Two approaches
- Hyptohesis and pavlues
- Confidence Interval estimation
Limitations of statistical inference
Only tells about the role of chance or random error in making inference from your study population to the source population.
- Does not tell about the role bias or confounding
- Statisitics do not tell you about causality
Bias
Systematic error in the design, conduct or analysis of a study that results in a mistaken estimate of an exposures effect on disease
Types of Bias
Selection Bias
Information Bias
Selection Bias
Systematic error in selecting subjects into one or more of the study groups such as cases and controls or exposed and unexposed
Information Bias
Errors in procedures for gathering relevant information
Ex. Bias in recall in collecting data interview
Confounding
Sitituation in which a non-causal association between a given exposure and an outcome is observed as a result of the influence of a third variable usually designated a confounding variable or confounded
A variable is a confounder if:
1) It is a known risk factor of the outcome
2) It is associated with the exposure but is not the result of the exposure
Is a covariate a cofounder
1) Is it associated with exposure
2) Is it causally associated with outcome
If YES then:
Calculate crude association
Calculate stat rum specific association
Confounding is not an _______ phenomenon
All or none