Introduction to Epidemiology

• Epidemiology is the study of health in human populations.
• Epidemiology is the study of how disease is distributed in populations.
• Epidemiology identifies disease causes and ways to control and prevent them.
• Epidemiologists study disease in groups, focusing on commonalities to infer principles.

Purpose of Epidemiology

• To understand the burden and causes of health problems.
• To make changes that decrease risk and improve health.

Historical Background: John Snow (1813-1858)

• Snow identified a cholera outbreak's source in 1854, and is considered an epidemiology founder.
• Snow argued cholera entered the body through the mouth, against the 'miasma' theory.
• Snow's ideas were published in an 1849 essay, 'On the Mode of Communication of Cholera'.
• In 1854, Snow identified a water pump in Broad Street as the cholera source.
• Removing the pump handle reduced cholera cases, but Snow's germ theory wasn't accepted until the 1860s.

Key Concepts

• Epidemiology focuses on the distribution and determinants of diseases
• The "3 D's" are Distribution, Determinants and Disease
• Frequency of disease occurrence varies among population groups
• Epidemiology examines the frequency and pattern of health events in a population.
• Frequency relates to the number of health events relative to the population size.
• Pattern refers to health-related events by time, place, and person.
• Determinants are factors or events causing a change in health.

Examples of Determinants

• Infectious agents
• Physical environment
• Genetics
• Lifestyle/behaviors
• Social environment
• Work conditions

Epidemiology questions about health events

• What is the event/problem?
• What is the magnitude?
• Where did it happen?
• When did it happen?
• Who were affected?
• Why did it happen?
• How did it happen?

Epidemiological Triangle

• The Epidemiologic Triangle is a model for studying health problems and infectious disease spread by applying a scientific model to different facts and circumstances.
• The triangle´s vertices:
  • Agent (what), microbe causing the disease
  • Host (who), organism harboring the disease
  • Environment (where), external factors enabling disease transmission
• Epidemiologists disrupt the connection between environment, host, and agent to stop disease spread.

The Agent - "What?"

• The agent is the cause of the disease

Examples of Biological Agents

• S. mutans cause dental caries
• Actinomuces causes root caries
• Vitamin C deficiency causes scurvy
• Herpes virus causes herpetic gingivostomatitis

Chemical and Physical Agents

• Chemical agents examples: lead, dust, medication, or fluoride (e.g., Fluorosis)
• Physical agents examples: radiation, heat, cold, or trauma (e.g., Head and neck radiotherapy cause salivary gland destruction and hence caries)

The Host - "Who?"

• A susceptible human or animal harbors a disease-causing agent
• Host susceptibility factors include age, sex, race, genetics, stress response, lifestyle, and immunization history.
• Some diseases are more common in males vs. females.

The Environment - "Where?"

• Encompasses all external conditions that influence the life and development of an organism/society.
• Physical environment: geography, climate, weather, water/air/soil (potentially with high fluoride, or high temperature with more water drinking)
• Biological environment: diet/nutrition, presence of animals/plants/insects, and protein/vitamin deficiencies during tooth formation affect tooth structure.
• Social environment: education, occupation, social conditions- all affect general health.
  • Educated people typically have better oral hygiene.

Time

• Most infectious diseases have an incubation period.
• Time can measure illness duration, or the time before death/recovery.
• Time marks period to epidemic threshold

Epidemiological Terms

• Endemic: Constant disease presence in a specific area (e.g., malaria in parts of the world)
• Epidemic: Unusual disease occurrence in a region
• Pandemic: Disease affecting many people worldwide
• Mortality: death rate (# deaths / # population)
• Morbidity: # diseased / # population
• Prevalence: existing cases of disease/health condition in a population at a set time
• Incidence: number of new cases

Rates

• Prevalence Rate: # Current cases / Population at risk
• Incidence Rate: # New cases / Population at risk

Susceptibility vs. Risk

• Susceptibility: Someone who can become a case is at risk.
• Risk factor: Exposure statistically linked to an outcome (e.g., smoking with periodontitis)

Risk Factor Criteria

• Disease frequency varies with exposure
• Exposure precedes disease onset (temporal sequence)
• Risk factors are modifiable or unmodifiable

Scope of Epidemiology

• Describe: Describe health status of populations.
• Explain: Explain disease origin (discover causality/transmission).
• Predict: Predict disease occurrence and distribution.
• Control: prevent new cases, eradicate existing, prolong lives via epidemiology.

The Uses of Epidemiology

• Study population health history.
• Diagnose community health in terms of incidence, prevalence, disability, and mortality.
• Study/improve health service operations.
• Estimate individual risks.
• Identify syndromes by describing the distribution and clinical phenomena in a population.
• Complete chronic disease picture and describe their natural history.
• Search for causes of health and disease.

Epidemiological Studies

• Study design choice is critical.
• Each design has strengths and weaknesses.

Types of Epidemiological Studies

• Observational: observes nature without intervention (descriptive or analytical)
• Experimental: attempts to change a disease determinant

Observational Studies

• Descriptive: describes occurrence of a disease in a population, and a first step in epidemiological studies
• Analytical: analyzes relationships between health status and other variables.

Experimental Studies

• Experimental or intervention studies involve an active attempt to change a disease determinant, and includes all clinical trial studies.

Observational Epidemiology

• Descriptive studies: A study where one group subjects who have the outcome (disease or any other healt related phenomena of interest), without a comparison group. Outcome or health- related phenomena are then are described according to the frequency of summary figures, and the distribution to selected variables related to person, place and time.
• Descriptive vs. Analytic Epidemiology
  • Descriptive answers who, what, when, and where.
  • Analytic focus on why and how.
• Descriptive provides disease condition description in a population
• Two main types: cross-sectional, and longitudinal studies

Cross-Sectional Studies

• Cross-sectional studies assess health and exposure simultaneously, a snapshot or 'prevalence study.'
• Advantages: easy to conduct.
• Disadvantages: can't infer causality; only studies current factors.
• Reveals existing cases or prevalence.
• Describes the disease occurrence.
• Simple, quick, inexpensive
• Data helps assess population healthcare needs.

Cross-Sectional Study Disadvantages

• Temporal sequence can't be established.
• Exposure-disease time cannot be mapped and measurements happen at same time.
• Cannot accommodate change and disease that take time to develop, and much info may be lost

Longitudinal Study

• Repeated examination of the same population over time
• Indicates new cases and incidence rate, describes natural history, identifies risk factors
• Can indicate efficiency of therapeutic measures

Analytical Epidemiology

• Studies the risk factor/outcome relationship
• Seeks health-related causes and effects and tests hypotheses.
• There are mainly two types:
  • Case-Control studies (retrospective).
  • Cohort studies (prospective)

Case-Control study

• Case-control studies, the "why me?" study, investigate individuals with particular health conditions and those without it.
• Cases become the "ill" and controls do not.
• Case-control studies study rare health outcomes without having to follow thousands of people, thus are quicker, cheaper and easier to conduct than the cohort study.
• Case-control studies begin w/the selection of cases which should be representative, and selected on the basis of disease.
• Controls are people without disease, matched to cases except for the presence of disease.
• An important aspect of case-control studies is the determination of the start and duration of exposure for cases and controls.

Selection of cases and controls

Case-control study characteristics:

• Assess how diseases relate to exposure/other variables
• Identifies cases and people who do not have the disease, and looking back to see if they had the exposure of interest.
• Typically, the study is retrospective.
• Used to study rare health outcomes
• Faster, cheaper than cohort study
• Can evaluate multiple risk factors.
• A major disadvantage is the potential for bias

Case-control study is subject to bias

• Bias: systematic/ non-random "playing favorites" in subject/exposure selection, invalidating the data
• Data and subjects must be recruited and collected uniformly.
• Selection bias occurs when not all eligible individuals can be selected. Need comparison populations
• Recall bias occurs when cases and controls recall exposures differently.

Addressing Recall Bias

• Verify exposure data via medical records and work histories.
• Use of controls with unrelated diseases, as they will likely consider prior exposures.

Cohort Study

• Cohort studies, also known as the "What will happen to me?" study, follow healthy people with varying exposures and notes health over time.
• Its desirable design is because exposure results occur before health outcomes; making the design less susceptible to bias.
• These studies can prove costly and time consuming.
• Cohort studies classify people by exposure to prospective disease/outcomes, measuring variables and monitoring new cases among groups with/without exposure.
• Cohort studies give best insights on disease causation, and most direct measurement of disease risk

Cohort Study Characteristics

• Can assess multiple outcomes
• Suitable for rare exposure
• Can estimate incidence rate
• Most useful for common diseases.
• Cohort study can be done in retrospective follow up to evaluate past exposure.

Retrospective Studies vs. Prospective Studies

• Retrospective studies provide a smaller sample size while requiring less time than prospective studies.
• Retrospective design is common for occupational epidemiology like mercury/amalgam assessment.

Clinical Trials

• Clinical trials (human testing of a drug) are set of procedure in medical research and drug development that are conducted to allow safety or more specifically information about adverse drug reaction.
• It is an experiment in humans that measures efficacy and safety through interventions in groups receiving a control intervention, concurrently.
• Group comparison tests are intended to test whether an agent alter the disease course favorably.
• There are two groups:
  • Test( receiving the agent under study)
  • Control (not subjected to the same agent under study).
• Both groups are similar through distribution of age, race, sex, socioeconomic status.

Effectiveness Versus Efficacy Trials

• Efficacy: tests an agent
• Effectiveness: tests agent under usual, everyday conditions after efficacy proven.
• Types of Clinical Trials: tests agent

Types of Clinical Trails

• Treatment trials: tests treatments, new drug combinations & surgical approachs
• Prevention trials: looks for ways to prevent or protect against recurring symptoms
• Diagnostic Trials: find test or procedure for diagnosing a particular disease or condition
• Screening trials: test best way to detect certain diseases and conditions
• Quality of life trials/supportive care: improve the quality of life for people with chronic illness
• Randomized clinical trials (RCTs):
  • Epidemiological experiment. Groups are randomly assigned, and divided into study and control
  • Group does or does not gets therapeutic intervention -Random allocation must be enforced

Components for Measuring Intervention

• Well-designed RCTs are measuring interventions
• RCTs are considered gold standard.

Validity Factors Include:

• Random allocation
• Use of control
• Blind concept
• Adequate test subjects
• Trial duration consistency
• Consistency in diagnosis

Random Allocation

• Each subject has equal "test or control" group chance
• Enable determining if intervention is causing determined outcomes What is control?
  • Comparison group has no exposure
  • Necessary to compare patient and control sets.

Types of Control

• Positive control: treatment of standard already set.

Additional Terms

• Negative controls: absence of active treatment or else a dummy treatment, called a placebo
• Placebo: Inert substance like active treatment.

Blinding Concepts

• Treatment awareness skews how results are recorded
  • Single Blind: Subjects don't know group assignments
  • Double Blind: Subjects/Investigators don't know assignments.
  • Tertiary Blind: The Statistician is unaware
• Clinical trials must be long enough for disease detection/lesion measurement

Time Frame of Results

• Caries trials (2-3 years)
• Plaque inhibiting agent (7-21 days).
• Gingivitis: (6 months)
• Calculus: (90 days)
• Must have adequate subject numbers.
• Loss of subjects can affect end results, must have large group sizes.
• Consistency/Reliability: Replicability of the results can be reproduce to the best known results

Consistency and Reliability

• Intra-examiner reliability: consistency within examiner
• Inter-examiner reliability: consistency between examiners
• Standardization calls for investigators training.

Advantages of RCT

• Comparative: One treatment is compared to another to establish superiority.
• Minimizes bias: Randomization minimizes allocation bias and selection bias.
• Bllinding minimizes performance bias.
• Randomization makes groups comparable to both known and unknown factors.
• Statistically reliable.

Disadvantages of RCT

• Might demand vast samples size, which require more resources from the investigators
• Allocation of participants may be predictable and result in selection bias when the study groups are unmasked.
• Longer trial duraton and more expensive
• May not mimic real life treatment situtations (e.g, inclusion/exclusion criteria; high controlled settings).
• Ethical limitations: some research cannot be ethically performed as an RCT.

Elements of Epidemiological Sampling

• When investigating a population, it is difficult to measure every individual
• Costly of time, manpower and funding
• A small sample/group must be selected
• Sampling allows researchers to gather data on a population without looking at every individual
• Sample chosen must be representative of entire populations

Common Sampling Methods

• Simple random sampling:
• Stratified sampling
• Systematic sampling
• Cluster sampling
• Convenience sampling
• Snowball sampling
• Simple random sampling is when individual is chosen chance
• All member of population has probability of selection

Important Facts of Sampling

• Stratified sampling population divided into similar.
• Select sample from each with purpose of creating similar subgroups
• This sample can change subgroups effectively and is good for all heterogeneous populations.

Systematic Sampling

• Individuals are selected at regular intervals from a list of the whole population.
• Intervals determined to ensure adequate sample size
• Saves time and is cheap to list large population features of high probability

Clustered Sampling

• Sub-groups sample the sampling units rather than individuals
• Chosen at random to select inclusion for study
• Geographic boundaries are commonly used.

Convenience Sampling

• Easiest method Investigator follows ways to chose in the study, such as an easy accessible road or neighbourhood

Snowball Sampling

• Ask new subjects to to nominate subsequent.
• Size expands similar to snowball.

Sampling size and factors

Sample Size

• Large sample gives more precise estimates.
• Factors affecting size:
  • Level of precision (need statistical power)
  • Homogeneity: how alike are its members
  • Prior information

Resources, Procedure, Geography, and Measurements

Resources effect

• Time, cost, personal will effect the sample.
• Sampling procedure
  • As the stratified sample requires fewer people than the simple random sample, the cluster sample requires more.
• Geography more dispersed; the population, the large sample size needed.
• Measurement, Types of questionaire is different

Bias & Confounding

• Epidemiological studies seek relationship links between exposure and validity, precision given resource limits.
• Validity: lacking bias and confounding.
• Bias: investigator error regarding study design in association with exposure
• Confounding reflects fact that epidemiological research done among free-living humans
• Precision is the association by mere chance.
• A study with a small systematic error is said to have a high accuracy.

Guidelines for Questionaire

• Once questions have been formulated, subjects identified its important that these subjects are recruited uniformly, that data about the health and exposure be collected consistently.
• If certain subjects aren't enrolled, invalid biased study

There are three types of broad categories of bias

• Selection bias
• Information/measurement bias.
• Confounding.

Selection Bias

• occurs when included subjects misrepresent
  • happens with small sample groups
  • Studies on volunteers often have this bias
• Occurs when difference are systematic

Measurement Bias

• Occurs when there's individual measurements, misclassification of inaccuracies, and/ or study variable inaccuracies due to subject error
• Instruments can also deliver bad date.

Recall Bias

• Retrospective cases may recall information for the wrong cases
• Cancer patient cases will try and attribute new symptoms for past occurrences or similar people
• Can prevent this by using verification of expsoure data, using medical records for drugs, diagnosis, and general records.

Observer Bias

• The investigator, laboratory technician or the participant knows the exposure status
• Prevented through blinding, where researchers cannot know the participants/ classifications

Instrument and Confounding Bias

• Instument, when its not properly calibrated, giving high or higher reads than expected. A complete picture gives the site and gives false information for distance
• Confounding
• In epidemological studies, confounding exists in population association for factors of being observed

Associations of confounding variable

• Confounding is difficult to control for and creates misleading estimations.
• The word "confounding" comes from the Latin 'confundere', meaning to mix together.
• Confounding has great influence
• It's important not to confuse with casual effect relationship
• It relates to a risk factor and associated by the exposure under investigation

How to controll in experimental studies

• In experimental studies
• Randomization
• Restriction
• Matching
• Randomization avoids potentially association.

Restriction

• One way to control confounding is people who have a restriction that have specific charcteristics.

Matching

• Match is used to ensure confounding by control group is similar

• Difference between bias and association is not the same truth

Validity

• Expression of how results/ test measure its truth.
• Study valid only if correspond of truth

Internal Validity

• Observation correct per specific group
• Dependent on amount of error
• Good internal means lacking measurement error and inferences drawn

External Validity

• Generatlizabiluty the extent which results study apply to people in it
• In the study supports more good results.

Causation and Epidemiology

• Major focus on prevention and control for disease/promote health

• Casualit/risk concept:

  • unfortunately in disease this is not always the concept
  • multi factorial effect if host/environmental factor is related

• Measures obtained for epidemiologist. studies= association or risk

• risk factor= raises risk of specific disease

Concept of Sufficient vs Necessary conditions

• Sufficient: An ineviatibility, a product or process that cannot happen without certain event.

Multiple Multi-factorial Conditions

• A sufficient cause can prevent actions.
• Smoking = lung cancer that's only partly attributable since others can influence genetic facts.

Component causes or Causal Pies

• Host-environement models don't account non-infectious diseases
• Rothman's Causal Pies model.
  • This model has come to be known as the Causal Pies After this, each factor is called a composite cause and pathways are sufficient. The last pathway for every pie are often neccessary.

Factors of causations

Categorized in four types.

• Factors include: predisposition, enabling/disbaling issues, precepitation, and reinforcement from harsh work due to poor wealth.

• The term risk factor describes factors of development that doesn't cause disease*

• Several diesases link coronary to heart but not the same factors

• Causation vs, association: Due multi- factorial disease, these are often dificult- determined

• Many risk factor exists with outcomes in individuals.

• Association: identiifiable relationship to expooure, implied to be cause.

Therefore desing, interventions.