Epidemiology
EpidemiologyEpi --------- upon Study of what comesDemos --------- people upon peopleLogos --------- study“Epidemiology is the study of the distribution and determinants of health related states or events in specified populations and the application of this study to the control of health problems"
Epidemiology has 3 main aims:
To describe the distribution and size of disease/ health problems in human population To identify aetiology and the pathogenesis of diseases 3. To provide the data essential for: management prevention evaluation for control of diseases planning of health services treatment /health problem To fulfill these aims, different classes of epidemiological studies are neededEpidemiological Studies
Observational Interventional (experimental) Descriptive Analytical Individual Group Group Individual Ecological Case report Case Cohort RCT Field Case series Control trials CSSSources of Data
A- Routine statistics 1. population statistics: e.g. Age, sex, geographical, mainly used for calculation of rate/ratio 2. Mortality statistics: e.g. Death registers 3. Morbidity statistics: e.g. notification of disease B- Surveys Mortality statistics: The simplest use is to provide Crude Death Rate (CDR) No of Deaths in a given period CDR = * 1000 population at riskStandardization
It is a process that permits comparisons among sets that show different compositions for factors like age and sex. Standardization helps to avoid biases that could arise as a result of differences in the composition of the population as regarding to age and sex. Standardization is aimed to calculate the standardized mortality ratio (SMR): SMR = observed deaths / expected deaths * 100 The expected deaths can be calculated by two methods: Direct (using a standard population) Indirect (using a standard ASDR)Population statistics
Morbidity statistics: - count events and not directly identify No. of people at risk - people entering hospital at number of times increase statistics patients spells - does not, in any way, provide definite answer to a particular problem, but only some background information.Types: Hospital discharge dataAbortion statisticsCancer registrationCongenital abnormalities statisticsInfectious diseasesMental healthRates and ratios
No. of New cases of a disease in the population during a specified period of time Incidence rate = *1000 population at risk of developing the disease during that period of time No. of all cases of a disease (new & old) in the population at a specified time Prevalence rate = *1000 population at risk of developing the disease at that period of time Prevalence rate is of two types: 1- Point prevalence 2- Period prevalenceLimitation of hospital statistics
Undiagnosed cases: (ice-berg phenomenon)Milder cases not reported ( treated elsewhere)3. Catchments area not defined Denominator is lacking rates can not be calculated4. Selection of patients 5. Deaths occur outside hospitalConfounding
DefinitionA situation in which an association between a given exposure and an outcome is observed as a result of the influence of a third unobserved factor, called a confounder.
Confounding
Is cause the only possible association between the two?1.
Cause
2.
Chance
3.
Bias
4.
Confounding
Cause
AssociationCause
X
Observed
Not Observed
Confounder
Toyota (Monika)
MI
Review - Diagram of Confounding
Fat intake
Epidemiological Studies
. Observational Studies – examine associations between risk factors and outcomes the researchers do not interfere or manipulate any of the factors under study. They record their observations of what’s going on, and explain what they observe with measures of associationsDescriptive - patterns and frequency of diseaseAnalytical - determinants and risk of disease Intervention Studies – explore the association between interventions and outcomes. researchers actually manipulate those factors they think have something to do with causing some outcome.Epidemiological Studies
Observational Interventional (experimental) Descriptive Analytical Individual Group Group Individual Ecological Case report Case Cohort RCT Field Case series Control trials CSSEpidemiological Studies
Time Past Present FutureA B C I A = Descriptive B = Case-control study C = Cohort study I= Intervention study
Descriptive studies
Shows prevalence of disease/condition Describe association with suspected risk factors Generate BUT NOT assess hypothesis Do not have a comparison group Traditionally focused on person, place &time Now focused on 5 Ws: Who ,Why, When, Where &WhatDescriptive on group
Ecologic(correlation) studySimplest typeInformation collected on group of people eg. schools, countriesOutcome-------support or not the hypothesisMay --------------show strength of associationGives very few information about confounding factors.Most cross sectional in nature…..Multi group studiesSome longitudinal in nature…….. Time series studiesLike CSS ….…distribution of particular outcomeLike CSS…… not suitable to test causal association.Descriptive on Individual
Case ReportCase seriesCross Sectional Studies(CSS) Cross Sectional Studies(CSS)They yield information which is of immediate relevance to the planning of medical services and to disease classification and natural history.Sometimes specific hypothesis e.g.More frequently indicate problems that demand further studies, e.g. variation in cancer incidence between countriesIn infectious diseases to understand inter relationship between environment, disease agent and human host.Cross Sectional Studies(CSS)
Principles involves in Cross-sectional studiesDefine aimChoice of study populationCategorizing data to be calculatedSampling why to do sampling? Not possible in - the only feasible way - data should be collected from total - financial reasons population - produce quick results - rare events - accuracy increase in certain surveys - sampling discrimination - response rate may be higher - 75% sample take 100%Sampling should be: Random and Representative5. Response rate
Advantages & disadvantages of Cross-sectional studies
Simple Quick Useful to quantify the health status of a population * Because done at a point of time: 1- unable (usually) to test causal relationship 2- extremely different to test the influence(s) of aetiological factors 3- (1 & 2) make cross-sectional study unable to test Hypotheses.Problems in Cross-sectional studies
Importance of accurate data: the data collected should be Reliable and Valid Reliability: extent of agreement between repeated measurement. This consist of the sum of variation in the item being assessed and the error introduced by the observer collecting information. Validity: is the extent to which a method provides a true assessment of that which it purposed to measure. 2. Introduction of biases 3. Accuracy of subjects responses 4. Accuracy of examination findings 5. Accuracy of investigation results 6. Bias introduced by non-response
s
Reasons behind conducting sampling include: presence of rare events.. financial. produce quicker results. getting higher response rate in some studies. being the only feasible way. Cross-sectional studies are: useful to quantify health status of a population. useful to test causal relationship.. easy. quick. unable to estimate incidence rate.
Case-Control Study
Often called Retrospective studies Are common first approach to test causal hypothesis Have three distinct features: Both exposure and outcome have occurred before the start of the study The study proceeds backwards from effect to cause; and It uses a control or comparison group to support or refute an inference.Case-Control Study
Study Population Time DISEASED non-DISEASED (Cases) (Controls)exposed non-exposed exposed non-exposed
Selection of cases and controlsSelection of cases Case definition Source of cases: hospitals, outpatient clinics and general population. Selection of controls Controls must be as similar to the cases as possible, except for the absence of the disease under study. Sources of controls: hospital, relatives, neighborhood, general population. *MATCHING
CASE-CONTROL STUDY
ADVANTAGES…Simple to conduct cost/time effective Able to look at multiple exposure at one timegood for studying rare diseases, and diseases with long latency period like cancers can use smaller sample sizesCASE-CONTROL STUDY
DISADVANTAGES…Highly subjective to bias (ex: selection and recall)can’t calculate incidenceNot good for rare exposuresselecting appropriate controls can be challengingClinical cases are selective survivors
Measuring the risk in Case-Control study
The incidence rate among the people exposed can’t be calculated, so we can’t calculate the risk directly. However an approximation of the relative risk can be derived which is termed the Odds Ratio “OR”OR = a*d / b*c OR=1 exposure is not related to the diseaseOR > 1 exposure is positively relatedOR < 1 exposure is negatively related Cases (Diseased)Control (not diseased)
Exposed
a
b
Not exposed
c
d
Cohort Study
Follow up study, prospective study, longitudinal study, incidence study.The term “cohort” is defined as a group of people who share a common characteristic or experience within a defined time period (e.g. age, occupation, exposure to a drug or vaccine, pregnancy, etc)The cohort are identified prior to the appearance of the disease under investigationThe study proceeds forward from cause to effectCohort Study
Time Study Population Free from the disease under study EXPOSED not EXPOSEDdisease no disease disease no disease
Cohort Study
Advantages: - Gold standard for studying the association between a risk factor and outcome - Useful for studying incidence, risk factors, natural history or prognosis - Useful for studying multiple outcomes - Useful for looking at multiple exposures and their interactions - No recall bias, less likely for selection bias
Cohort Study
Disadvantages:- Expensive- Often Long-time for follow-up - Large sample size need- Not good for low-incidence (rare) diseases- Not good for chronic diseases with long latency- Attrition – “loss to follow-up”Measuring the risk in Cohort study
As cohort study provide the data needed to calculate the incidence rate of the disease among exposed and among the non-exposed, so we can calculate the Relative risk “RR” and Attributable risk “AR”RR= Incidence in exposed / incidence in non exposed RR = a / a+b / c / c+dRR = 1 : No evidence for any association bet exposure and risk of disease RR > 1 : evidence of positive association and may be causalRR< 1 : evidence of negative association and may be causalAttributable risk: the amount or proportion of disease incidence (or disease risk) that can be attributed to a specific exposure. AR= Incidence in exposed - incidence in non exposed AR= a / a+b - c / c+d Attributable Risk Percent (AR%) : AR% = AR / incidence among exposed * 100 = AR / a/a+b * 100
Interventional (Experimental) Studies
These studies involve an active attempt to change a variable in one or more groups of people. Same design as a cohort study with one vital difference, that the exposure status of the study population has been deliberately changed by the investigator. We observe how this change in exposure alters the incidence of disease or other features of the natural history. They have all the advantages and disadvantages of the usual prospective cohort studies plus the additional problems of cost and ethics.Interventional Studies
Individual..... (Randomized Controlled Trial) (RCT) Group………(Field trial)The Randomized Controlled Trial (RCT)
the ultimate study design; the "gold standard" against which all other designs are compared. The subjects are usually chosen from a large number of potential subjects using a set of inclusion and exclusion criteria. an informed consent is obtained from each participant. Randomization is then done to allocate subjects to either the treatment group or the placebo group. Once randomization is done, intervention is begun. After the intervention, the key outcomes that are being studied need to be measured by and analysis involves looking for differences in the outcome rates in the two arms of the trial.Time
CHDCHD
CHD
CHD
CHD
now
I
I
I
I
I
I
I
I
I
I
I
I
I
C
C
C
C
C
C
C
C
C
C
randomly allocate
Control group
Intervention group
future
Intervention studies
RANDOMIZATION:It is the heart of an RCTIt is an attempt to eliminate “bias” and allow for comparability.It ensure that the investigator has no control over allocation of participants to either study or control group thus eliminating what is known as “selection bias”BLINDING:Is done to ensure that the assessment of outcomes is done objectively away from biasSingle blind trialDouble blind trialTriple blind trial
Field trials: in contrast to clinical trials, involve people who are disease-free but presumed to be at risk Data collection takes place in the field because probability of disease is small, a large number of subjects are needed One of the largest field trials ever undertaken was that of the Salk vaccine for the prevention of poliomyelitis, which involved over one million children.
Validity of measurements
Validity is the sensitivity and specificity of a test Sensitivity: is the ability of the test to diagnose correctly the positives out of the total who should be positives Specificity: is the ability of the test to diagnose correctly the negative out of the total who should be negativeValidity of measurements
Sensitivity [a / (a+c)] Specificity [d / (b+d)] Predictive value of a positive test (PPV) [a / (a+b)] Predictive value of a negative test (NPV) [d /(c+d)]Gold Standard
Disease
+
No disease
-
Test
a
b
c
d
+
-
Validity of measurements
Sensitivity = [95 / (100)]=95% Specificity = [50 / 100] = 50% Predictive value of a positive test (PPV) = [95 / 145] = 65.5% Predictive value of a negative test (NPV) = [50 /55] =91%Typhoid Fever
Culture
+
Culture
-
Widal test
95
50
5
50
+
-
Likelihood Ratios
The likelihood ratio for a positive test result (LR+) is the probability of a positive test result for a person with the disease of interest divided by the probability of a positive test result for a person without the disease. LR+=Sensitivity/1-specificity An LR+ of one indicates a test with no value in sorting out persons with and without the disease of interest The larger the value of the LR+, the stronger the association between having a positive test result and having the disease of interest.The likelihood ratio for a negative test result (LR–) is the probability of a negative test result for a person with the disease of interest divided by the probability of a negative test result for a person without the disease. LR– = 1-sensitivity/ Specificityvalue of one indicates a test with no value in sorting out persons with and without the disease of interestThe smaller the value of the LR–, the stronger the association between having a negative test result and not having the disease of interest.in contrast to predictive values, the likelihood ratios do not vary according to the prevalence of disease.