Epidemiology
epidemic
disease excess of expected, differs by type of disease
epidemic curve
distribution of the times of onset for cases occurring during an outbreak -Tells about mode of transmission, occurrence of an epidemic, time and source of exposure, mode of transmission, causative agent
ratio
divide one quantity by another, no specified relationship between numerator and denominator. Example; 1000 motorcycle fatalities 950 are men, 50 women 950/50= 19:1 male to female. It's a general term that includes specific measures such as proportion, percentage and rate
indirect adjustment
does not require knowledge of actual age-specific incidence or mortality rates among each group. •Comparison of one population group's rates against a standard population's rates using the first group's base population. •Often compares a group at special risk against the basic group from the whole US •Doesn't require knowledge of actual age-specific incience or mortality rates among each age group.
direct adjustment
equalize differences in population distribution by applying an outside, standardized population to the rates of the groups to be compared. The standardized population may be: a combination of the 2 populations, one of the 2 populations, a completely outside population. •Compares 2 study populations. Adjusted rate for each study population is determined by normalizing to a standard population. Means nothing on its own. Requires a comparison. •Commonly called standardized. CRUCIAL TO MENTION THE STANDARD!
paired percent agreement
measure of the agreement between two tests/obervers divided by the total number of pairs where at least one of the two tests/observers indicate that a positive result has occurred.
overall percent agreement
measure of the total agreed upon number of observations between two tests/observers out of the total number of observations.
cross tabulation
method for examining the influence of multiple variables in combination
reportable disease statistics
usually infectious and communicable diseases Advantage: track outbreaks of infectious diseases Limitations: incompleteness of population coverage, physician failure to fill out reporting forms
Census Data
valuable source of information - social, economic characteristics Advantages - complete data Limitations - some segments of population may be undercounted.
expected agreement
determined by adding together the number of positive test results that we expect the observers to agree upon by chance along and the number of negative test results that we expect to be agreed upon on chance alone.
trade-offs
determining a cut-off level where a test result is considered positive or negative is needed. Depends on nature of disease. Sometimes you decide cutoff.
1990s
- application of techniques in molecular biology to large populations
disease incidence
# of NEW cases of a disease occurring during a specified period of time in a population at risk for the disease; describes the rate of development of a disease in a group over a certain period of time. *AT RISK if they DO NOT HAVE IT
negative declaration
(null hypothesis): there is no difference between the infant mortality rates of 2 regions.
positive declaration
(research hypothesis): the infant mortality rate is higher in one region than another
horizontal transfer
(transmission within a population, between a source and vulnerable person) vs. vertical (from other to baby, of a genetic or infectious nature)
John Snow
- "Father of Epidemiology" -Investigated cholera 1849-1854 -Lambeth company - water above London -Southward and Vauzhall Company - water below London -1854 epidemic of cholera in London -Tested hypothesis -Charted frequency and distribution of disease -Ascertained cause/determinant -First to draw all three components of epidemiology together oDistribution, determinants, population -Important: it is not always necessary to know pathogenic mechanism to prevent disease
Salk Vaccine
- 1954 - largest formal human experiment
Functions of Epidemiology
- To discover the agent, host, and environmental factors which affect health, in order to provide the scientific basis for the prevention of disease and injury and the promotion of health. - To determine the relative importance of causes of illness, disability, and death, in order to establish priorities for research and action. - To identify those sections of the population which have the greatest risk from specific causes of ill health, in order that the indicated action may be directed appropriately. - To evaluate the effectiveness of health programs and services in improving the health of the population.
1964
- US Surgeon General's Advisory Committee on Smoking and Health
life tables
- a way of using the actual observed survival over time to measure prognosis. Survival rates/percentages can be determined for specific time intervals or aggregate periods (resulting in cumulative survival probabilities). - The probability of surviving any particular year is equal to the number of people alive at the end of that year divided on the number who were followed for that year. Assume that: oNo changes in survivorship have occurred over calendar time. oThose lost to follow-up experience the same survivorship as those who are followed for the full period. Issues: oImprovements in diagnostic methods over time oImportant to account for changes in treatment effectiveness before assuming improved survival.
1970's
- evolution of microcomputer technologies allowed new multivariate statistical methods to develop
adjustment of rates
- makes comparison of morbidity or mortality rates between two groups more accurate (indirect or direct)
Case-Fatality Rate
- the number of persons who die from a disease within a certain period of time divided by the total number of people who have the disease
Bacteriological Revolution
-"Point-contact spread" of infection
Two by Two for a screening test
-A = True Positives (TP) - have disease and have a positive test -B = False Positives (FP) - no disease but have positive test -C = False Negatives (FN) - have disease but have negative test -D = True Negatives (TN) - no disease and have negative test -Sensitivity = (A) / (A+C) OR (True Positives) / ((True Positives) + (False Negatives)) -Specificity = (D) / (B+D) OR (True Negatives) / ((False Positives) + (True Negatives))
Population-based approach/public health approach:
-Applied to whole population, dietary modification, must be inexpensive and non-invasive
Develop and Test Hypothesis
-Based on knowledge of cases and disease formulate hypothesis -Using data collected test hypothesis (use measures of risk)
Calculate incubation period
-Collect patient data - case finding -For example, determine that 100 people who ate egg salad at a potluck lunch got sick -We find these people and ask them: -What time did you eat the egg salad? -What time did you feel the first symptoms? -Difference is incubation period -Calculate difference for each person and plot histogram
Control Measures
-Control outbreak -Prevent future outbreak -Timing important and difficult -Challenge to balance responsibility to prevent further disease with "political" considerations Ex: Credibility and reputation of an institution
Steps in Investigation of Disease Outbreak
-Determine if outbreak exsits -Distribution of cases -Look for combination of relevant variables -Develop hypothesis -Test hypothesis -Implement control and prevention efforts -Communicate findings
person to person contact
-Epidemic curve shows an epidemic that grows as it spreads -Propagated outbreak -Cases occur over >1 incubation period -Expect to see successive peaks reflecting increasing number of cases in each generation -Few actually show classic pattern
High-risk approach (clinically based):
-Expensive or invasive (ie. colonscopy for those with family history, cholesterol screening for children from high risk families)
Why Quantify Natural History of Disease?
-Help answer patients questions regarding prognosis, describe the severity of disease to establish priorities, establish baseline to evaluate new treatments, provides knowledge in assessing outcomes after the comparison of two treatments.
Epidemiological Approach
-Observation -Definition of disease process -Descriptive epidemiology -Analytical Epidemiology -Experimental Epidemiology
Communicate Findings
-Report to state and local health agencies, CDC -Inform others - businesses, manufacturers, etc. (ex. Tylenol) -Need to share information with public must be assessed (ex. anthrax)
many stages of disease
-Stage of susceptibility -Subclinical disease (before we recognize symptoms) -Clinical disease (mind, moderate or severe?)
Outbreaks - Epidemics
-The occurrence of disease in excess of what is expected -Unexpected - immediate response may be demanded -Fieldwork needed -Limited investigation because timely intervention needed -Use surveillance system to determine what is expected -Is disease endemic? -Is disease process known?
attack rate
-Type of incidence measure -Number of people at risk for a disease who develop the disease compared to the total number of people who are at risk for the disease
Issues in Epidemiology
-Where there is a link between factor and health outcome - does this mean the factor is the cause of disease? -Is there is an association, does the amount of disease vary according to the amount of exposure to the factor? -Based on the observation of such an association, what practical steps should individuals and public health departments take? -Do the findings from an epidemiologic study merit panic or measured response? -How applicable are the findings to settings other than the one in which the research was conducted?
Two assumptions inherent in epi studies
1. Diseases are NOT randomly distributed in populations. 2. Diseases have specific causes that can be identified, prevented, and treated.
years of potential life lost (YPLL)
A measurement of time lost due to mortality. 16/18yrs old through 65 yrs old are considered "productive" years. It is a way of measuring impact on a nation or society from premature loss of members due to a disease. Calculated at 65 minus age at death.
Objectives of Epidemiology
1. To identify the etiology (cause) of a disease and the risk factors associated with the disease. 2. To determine the extent of disease found in the community. What is the burden of disease? 3. To study the natural history and prognosis of the disease 4. To evaluate both existing and new preventative and therapeutic measures and modes of health care delivery 5. To provide the foundation for developing public policy and regulatory decisions relating to epidemiologic problems
Royal Society of London
1662 - John Graunt published a comparative study of mortality and morbidity in human populations. Referred to as the Columbus of Statistics. Quantified patterns of disease.
James Lind
1747 - etiology and treatment of scurvy, use of comparison group. -Noticed high morbidity rate among sailors -Noticed diet among sailors at sea hard on digestion -Conducted experimental trial using comparison group
Edward Jenner
1768 - use of observational data
Food Borne Disease Outbreak
2 or more people get sick from a common food
Hippocrates
400 BC - Found that disease is associated with physical environment
attack rate formula
= # of people AT RISK who develop illness / Total # of people AT RISK
secondary attack rate formula
= # of people exposed to a primary case who develop illness / Total # people susceptible after first wave
incidence rate formula
= (# new cases of disease occurring in population during specified time / Total population at risk of developing disease during that time) X per 1000 people
rates
A ratio with a distinct relationship between numerator and denominator and a measure of time is an intrinsic part of the denominator. Example; number of clods per 1000 elementary school students during a one month period.
Special Clinics/Hospitals
Advantages: specialized information on disease Limitations: cannot be generalized readily to reference population (exception: mayo clinic)
Physician Practices
Advantages: valuable supplemental information, verify self-report against medical records, source of exposure data Limitations - limited application, code of confidentiality and privacy, highly select group
Successes in Epidemiology and Public Health
Cholera, smallpox, Legionnaire's Disease, Infant Mortality/life expectancy, Toxic Shock Syndrome, smoking/tobacco (lung cancer), CHD - Coronary Heart Disease
factors in a screening test
Disease Related Factors: diseases often include a sub-clinical period, early treatment provides benefits to patients. Test Related Factors: valid and reliable test is available, test is inexpensive, easy to use, minimally intrusive Feasibility: willingness of the population to submit to screening, treatment is available if disease is detected. -Serious (breast cancer), Pre-symptom treatment should be more beneficial than if given after development of symptoms, prevalence of preclinical disease should be high in screened population (target screening) Example: Hypertension - serious, risk of death increases with higher levels of blood pressure, early treatment reduces risk
Teenage smoking example
Host = teenagers; agent = cigarettes; vector/vehicle = advertising; environment = social setting
Uses of Surveillance Information:
Immediate: epidemics, emerging health problems, changes in health practices, changes in antibiotic resistance Annual: estimating magnitude of health problem, assessing control activities, setting research priorities, testing hypotheses, etc Archival: describing natural history of diseases, facilitating epidemiologic and lab research, setting research priorities, documenting distribution and spread
Francis Bacon
Inductive logic, Law of Mortality --> law of epidemics.
Epidemiology
It is the study of the distribution and determinants of health related states or events in specific populations and the application of this study to control health problems.
multiple exposure
More than one contact (TB)
William Farr
Mortality surveillance, use of vital statistics, -In charge of medical statistics in the Office of the Registrar General for England (1839) -Set up a system for routine compilation of the numbers and causes of deaths -Addressed many issues relevant to modern epidemiology including case definition, using comparison populations, and addressing confounding factors such as age.
SEER Program
National Cancer Act 1971 established the National Cancer Program under which the Surveillance, Epidemiology, and End Results Program was developed. SEER collects cancer data throughout the US.
Surveillance
Ongoing and systematic collection, analysis and interpretation of health related data -Assess public health status, define public health problems, evaluate programs, stimulate research
Three Stages of Modern Epidemiology
Sanitary statistics, Infectious disease epidemiology, Chronic disease epidemiology
Positive Predictive Value
The likelihood that someone with a positive test result really has disease. Can be determined by using a gold standard as comparison.
persistent (chronic) infection
a chronic infection with continued low-grade survival and multiplication of the agent
latent infection/disease
an infection with no active multiplication of the agent. In contrast with a persistent infection, only the genetic message is present in the host, not viable organisms.
subclinical
an infectious with no clinical symptoms, usually diagnosed by a serological response or culture. People with unapparent infections are often able to transmit the infectious to others.
vector
any insect or living carrier that transports infectious agent from infected individual or its waste to a susceptible individual or its food. -Ex. Ticks carry rickettsiae, which cause Rocky Mountain Spotted Fever
clinical phase
any time after disease onset where signs and symptoms are present. Defining the Disease: must determine when disease starts, case use histological confirmation, screening as a disease locator, must determine stage of disease
preclinical phase
anytime after disease onset while there are no outward signs and symptoms
screening
application of a test to people who are as yet asymptomatic for purpose of classifying them with respect to their likelihood of having a particular disease. Risks and costs must be weighed against benefit. Not diagnosing illness.
choosing cut-offs
ask yourself: What is the consequence of an undetected case? What is the consequence of calling a disease-free person a positive? At what level does risk increase (i.e. at what point should someone receive treatment?) *Sensitivity and Specificity wont change, they are the same for every population regardless of disease prevalence. If you change the cutoffs at which you define disease, the sens and spec DO CHANGE.
observed survival person-years
assumes that person-years are equivalent. For example, the experience of 1 person observed for 10 years is the same as 10 people observed for 1 year each.
birth certificates
calculate birth rates, birth conditions, studies of environmental influences on congenital malformation. Advantages: complete, includes other information Limitations: some data unreliable, some conditions undetected at birth
disease registries
centralized database for collection of information about a disease Advantages: study incidence of cancer or selected disease, used in selecting cases for a case-control study, complete for disease Limitations: dependent on funding and staff, non-reporting bias
Relative Survival
compares the survival of a group to the survival we would expect in the group if they did not have the disease. Greater in older age groups.
Pierre Charles-Alexandre Louis
comparison of groups of individuals, emphasized use of statistical methods in medicine
continuous exposure
constant and always present (cholera)
Vehicle
contaminated inanimate object that transmits the disease (ex. doorknob)
case definition
criteria that must be met for the patient to qualify as having the disease (based on clinical symptoms or test results)
Sources of Mortality Data
death certificates, mortality reports/surveillance Problems - classification of cause of death, underlying vs. contributory cause, accuracy of causes of death, population data
incidence rate/incidence density
denominator consists of the sum of different times each individual was at risk -- expressed as person-time (cases per person-years)
active surveillance
depends on periodic solicitation of case reports from health care providers or facilities. Generally more accurate and complete but requires more effort and is expensive.
kappa statistic
evaluates how much better then random chance is the agreement between two tests/observers. Observed agreement is equal to overall percent agreement. Normally a value between 0 and 1 - How many positive test results do we expect the observers to agree upon? -How many negative test results do we expect the observers to agree upon?
Gold Standard
external source of truth regarding the disease status of each individuals. Defines the presence of disease. The best diagnostic assessment available. May be another test.
index
first case in a family or other group to come to the attention of the investigator
Infectious disease epidemiology
germ theory (single agent, specific disease)
endemic
habitual presence of disease within an area (ex. malaria in Africa, Lyme disease in Northeastern USA)
Predictive Values
highly affected by disease prevalence and specificity and lesser so by sensitivity. The PPV in one population will not be the same as in another population. PPV decreases with decreased prevalence and specificity and increases when prevalence and/or specificity is high. NPV decreases when prevalence increases and/or specificity is low and vice versa.
carrier status
individual that has the organism but is not infected, capable of transmitting
hospital and clinic records
inpatient and outpatient data Advantages: centralized data, information on hospitalized condition, useful for selecting cases and controls Limitations: individuals don't represent any specific population, lack of standardization in type of information collected on each patient, completeness depends on physician
Secondary prevention
intent to reduce progress of disease, screenings for "high risk" subjects - detection. Involves people who already have a disease.
implicit question
is there an association between infant mortality and geographic region of residence? →Impacts how we move forward
Kaplan-Meier
life table approach uses intervals to group events. Kaplan and Meier developed a method does not use intervals but calculates the survival probability each time an event occurs.
preclinical
not yet apparent, but will progress
Point Prevalence
number ill at a point in time divided by the total number in group at a point in time.
Period Prevalence
number ill during a time period divided by the average population during a time period.
count
number of cases, number of deaths
proportion
numerator is part of the denominator, may be expressed as percentage. Example; number of fetal deaths over the total number of births.
single exposure
one time contact, all infected at same time, single epidemic peak (church dinner)
School Health Programs
physical exams, immunization history, cognitive and other tests Advantages: uniqueness of data, school-aged population, potential for years of data on individuals Limitations: sporadic/incomplete data
Sanitary statistics
poisoning by soil, air, water
nonclinical infection
preclinical, subclinical, persistant, latent
clinical infection
presence of signs and symptoms
Crude Rates
presented for entire population, summary measures, actual # of events over a given time period. The proportion of population that dies during a time period.
Cumulative/Lifetime prevalence
prevalence of a disease during the lifetime of an individual. (have you ever had syphilis, how many times have you fallen?). Equivalent to a lifetime incidence, especially if length of the lifetime is accounted for. However, it is highly subject to recall bias.
Primary prevention
prevention or cessation of risk factor exposure (active or passive)
Why Look at Mortality Data?
provides clues to changes in patterns of disease occurrence. Mortality can be reflection of incidence when case-fatality rate is high, duration of disease is short.
single source, single exposure
rapid rise in cases, cases are limited to those who share the exposure, secondary cases are rare
Tertiary prevention
reduces limitation of disability from disease.
sentinel surveillance
relies on reports of cases of disease whose occurrence suggests that the quality of preventive or therapeutic medical care needs improvement. Serves as warning to health officials. Inexpensive. Lacks specificity regarding cause of disease and risk factors (used for example if there is a case of polio).
passive surveillance
relies upon reporting of cases by health care professionals at their discretion. Routine reporting. Prone to error. Inexpensive.
"Withdrawn" people
remove them completely from being at risk, count them completely as being at risk, count them for being at risk part of the interval.
series testing
sequential testing, we have two screening tests, someone is defined as having disease if they test positive on BOTH, if someone tests negative on EITHER test they are considered to not be diseased.
disease cluster
several cases of disease in a certain area
Chronic disease epidemiology
since WWII, "black box" approach, risk factors
case
single person, individual with illness
morbidity data
surveys collect data on health status of a population Advantages: representative sample, data quality, many variables, insurance data Disadvantages: insurance data not representative, may be over-representative of healthier individuals
diagnostic and/or screening tests
tests designed to separate persons without disease from individuals with disease based on biological variable. Ex: Diabetes (blood sugar)
validity
the ability of a test to accurately indicate which individuals in a population have a disease and which do not.
specificity
the ability of a test to correctly identify individuals who do NOT have the disease.
sensitivity
the ability of a test to correctly identify those who have the disease.
secondary attack rate
the attack rate in susceptible people who have been exposed to a primary case.
immunity
the capacity of a person when exposed to an infectious agent to remain free of infection or clinical illness
cumulative incidence
the incidence calculated when all persons within the denominator are considered to be at risk for the disease for the same period of time.
Median Survival Time
the length of time until exactly half of the study group has died. Not effected by outliers. Can be determined sooner (mean survival time requires that all deaths have occurred).
Negative Predictive Value
the likelihood that someone with a negative test result really does NOT have disease. Can be determined by using a gold standard as comparison.
reliability
the measure of whether a test gives the same results after repeat testing. Use: overall percent agreement, paired percent agreement, kappa statistic
prevalence
the number of EXISTING CASES of a disease or health condition in a population at some designated time. Time can be short or long; =incidence x duration of disease --useful for determining the extent of disease or health problems, and for assessing the public health impact. It may be used to estimate the likelihood of having disease in population
Specific Mortality Rate
the number of deaths during a set period divided by the mid-point population limited to a specific disease, population group or combination.
Mortality
the number of deaths occurring during a period of time within a specified population. Mid-point populations often used in calculating rates. All persons in the pop. must be at risk for death.
case-fatality rate
the number of persons who die from a disease within a certain period of time divided by the total number of people who have the disease (%) - best used when disease is short term, not chronic.
direct age adjustment
used if age-specific death rates in a population known and suitable standard population is available. Requires application of observed rates of disease in a population to some standard population to derive an expected number of mortality.
Surveillance
the ongoing systematic collection, analysis and interpretation of outcome-specific health data, closely integrate with the timely dissemination of these data to those responsible for preventing and controlling disease or injury. Monitors changes in disease frequency, prevalence of risk factors, changes in environment, vaccination coverage. 3 types: passive, active, sentinel
Proportionate Mortality
the percent of deaths attributable to a specific cause of death (not a rate, can be specific to demographic in conjunction to disease)
5 year survival
the percentage of people alive 5 years after diagnosis/initiation of treatment.
herd immunity
the resistance of a group to attack by a disease because a majority of individuals are immune; thus lessening the likelihood that a susceptible person will come in contact with a patient with the disease -Is it necessary to vaccinate everyone? -Not necessary to achieve 100% immunization, immune large percentage and the rest will be immune. Optimal when populations are constantly mixing together.
incubation period
time we are first exposed until first symptoms, helps determine the cause of disease because diseases have different incubations periods
parallel testing
two screening tests. Someone is defined as having disease if they test positive on EITHER test. Someone has to test negative on BOTH tests before they are considered to be free of the disease.
pandemic
worldwide epidemic (Ex. 1919 Flu, obesity, HIV, chronic diseases)
Direct vs. Indirect Adjustment
•Indirect adjustment compares a study population to a standard. •Can be used in any situation where a standard has been determined (cheap and easy). •Small random error (standard rates are determined from very large samples) •Two SMR cannot be compared. Only the mortality in the study group and the rates in the standard population.
incidence/prevalence issues
•Numerators: defining who has the disease •How do you find the persons with the disease? •Problems with recording data/ hospital data oHospital admissions are selective oHospital records are designed for treatment/administration, not study oDenominator- what population is represented by cases identified at a hospital? •Denominators- undercounting of minority groups •Definition of persons at risk