Epidemiological Studies PDF

EPIDEMIOLOGICAL STUDIES Types of epidemiological investigation Descriptive – Those undertaken without specific hypothesis. – Often the earliest studies done on a new disease in order to characterize it, quantify its frequency, and determine how it varies in relation to indivi...

EPIDEMIOLOGICAL STUDIES Types of epidemiological investigation Descriptive – Those undertaken without specific hypothesis. – Often the earliest studies done on a new disease in order to characterize it, quantify its frequency, and determine how it varies in relation to individual, place and time. Analytical – Analysis of observation using statistical and diagnostic tests – Undertaken to identify and test hypotheses about the association between an exposure of interest and a particular outcome.  Types of epidemiological investigation Theoretical – Consists of representation of disease using mathematical models that attempt to simulate natural patterns of disease occurrence. Experimental – Analyzes data from group of animals from which he can select and from which he can alter the factors associated with the groups. – The investigator has the ability to allocate animals into categories – also designed to test hypotheses between specific exposures and outcomes — the major difference is that in experimental studies the investigator has direct control over the study conditions Types Examples Descriptive Routine data, case reports, case series, field surveys, prevalence surveys Analytical Case-Control studies Cohort studies Cross-sectional studies Theoretical Mathematical modelling Prediction studies Experimental Clinical trials Field trials Community trials  Where is the disease especially common or rare, and what is different about those places? Usually better presented pictorially in a map How does disease frequency change over time, and what other factors are temporally associated with those changes?  Descriptive Studies Routine data Health and productivity profile Case studies/reports Case series Field surveys (ex KAP) Prevalence surveys  Routine data Examination of data routinely collected and submitted by various animal facilities – Slaughterhouses – laboratories Health and productivity profile  Case Report Case series – collection of case reports  Field Survey Locally known as KAP survey- “KAP” for knowledge, attitudes and practices  Prevalence Survey Cross-sectional surveys Selection of a random sample of animals from a population and the disease status and exposure status (potential risk factors) are measured simultaneously  Analytical Epidemiology Explanatory observational studies Is aimed at determining the strength, importance and statistical significance of epidemiological associations. comparison of two or more groups is the foundation of this design  Analytical Studies Comparative studies testing a hypothesis 1. cross-sectional a snapshot; no idea on cause-and-effect relationship 2. cohort prospective; cause-and-effect relationship can be inferred 3. case-control retrospective; cause-and-effect relationship can be inferred Are exposures linked with the disease? E D Exposure Disease Hierarchy of Evidence Gundran, 2015  Nomenclature of Analytic studies Cross-Sectional Longitudinal Case control Cohort Synonym: Synonyms: Synonyms: Prevalence Retrospective Prospective Case comparison Incidence Case referent Longitudinal Follow-up Analytic Studies (observational)  Cross-sectional studies Selection of a random sample of animals from a population and the disease status and exposure status (potential risk factors) are measured simultaneously useful for describing the disease situation at the time of data collection and it allows determining prevalence. Cross-sectional study Question: What is happening?  Cross-sectional studies Often used to study conditions that are relatively frequent with long duration of expression It measures prevalence, not incidence of disease It is easy, quick and inexpensive Example: community surveys  Cross-sectional studies Often used to study conditions Advantages that are Disadvantages relatively frequent with long duration of -When random sample is selected, -Unsuited to the study of rare diseases expression disease prevalence and proportions -Unsuited to the study of diseases of It measures in exposed and unexposed prevalence, in target short not incidence of duration population can be estimated disease -Control of extraneous variables -Relatively quick to mount and maybe incomplete Example: community-Incidence conduct surveysin exposed and unexposed -Inexpensive individuals cannot be estimated -No risk to subject -Temporal sequence of cause and -Current records can be used effect cannot necessarily be -Allow study of multiple potential determined causes of disease  Case-control studies study of persons/animals with the disease (“cases”) of interest and comparison group of persons/animals without the disease (“controls”) Frequencies of suspected factors are then measured in two groups  Case-control study Note: you start with diseased animals Question: What happened?  Case-control studies Advantages Disadvantages -Well suited to rare diseases -Exposed and unexposed -Quick proportions in target population -Inexpensive cannot be estimated -Few subjects -rely on recall or records for -No risk to subjects information on past exposures -Allow study of multiple potential -Validation of information is difficult cause of disease or sometimes impossible -Selection of an appropriate comparison group may be difficult -Incidence in exposed and unexposed individuals cannot be estimated  Cohort Studies A cohort (group) of healthy animals capable of developing a disease of interest is divided into two groups It selects groups according to the presence or absence of exposure to the hypothesized factor Followed (monitored) forward for a period of time the frequency of disease of interest is compared between groups  Cohort Studies Can either be prospective or retrospective the most effective observational study for investigation of causal hypotheses with respect to disease occurrence  Prospective cohort study Note: you start with exposed animals Question: What will happen?  Cohort Studies Advantages Disadvantages Often called prospective (concurrent study) -Incidence in exposed and -Exposed and unexposed target theindividuals unexposed most effective can be observational population cannotstudy for be estimated investigation of causal calculated hypotheses -Large with are numbers of subjects respect -well suited to disease for studying rare occurrence required to study rare diseases exposures -Long duration -Permit flexibility in choosing -Expensive variables to be systematically -Difficult follow-up recorded Retrospective Cohort Study  Example of Cohort Study: o Select a population of pregnant cows who took a drug and some who did not o Observe group over time to determine how many exposed and non-exposed calf developed malformation o Then now we can compare the proportion of malformed calf in the exposed group and the proportion of malformed calf in the non-exposed  Criteria Cross- sectional Case-control P. Cohort Sampling Random sample separate separate of the study samples of samples of population diseased and exposed and non-diseased non-exposed units Unit Time One time Usually Follow-up over a retrospective specified period Measure of Prevalence Prevalence of Incidence density frequency exposure and cumulative incidence Measure of Odds ratio Odds ratio Odds ratio association Relative/ risk Relative risk ratio Experimental study  Field trial A field trial is a comparative study involving new treatments or preventive measures applied under natural, field or semi-field conditions. Measure of Association  Concept of risk RISK : that it is the probability of an untoward event. Risk factors include any factors associated with an increased risk of becoming diseased or to die Any investigation into the cause-effect relationships between potential risk factors and an outcome parameter such as disease or death involves calculation of risks.  Factors that interfere with assessment of risk Long latency period/incubation High prevalence of risk factors/ disease Low incidence of disease Multiple causes Small risk from exposure  Uses of risks Prediction – presence of risk estimate likely future incidence of disease among comparable individuals Diagnosis – Presence of risk factor in an individual increases the likelihood that an associated disease is present Cause identification – Not because risk predict disease, it does not follow that it cause disease. A risk factor may mark a disease outcome indirectly, by virtue of an association with some other determinant of disease – A risk factor that is not a cause of disease is called a “marker” because it marks the increased probability of disease Prevention – If a risk factor is the cause of disease, its removal can be used to prevent disease even if disease mechanism is unknown. Comparing risk in cohort studies To compare risk, several measures of the association between exposure and disease, called measures of effect, are commonly used. They represent different concepts of risk and are used for different purposes  Measures of effect Expression Clinical Question Calculation Relative Risk (risk ratio) How many times more likely are RR = IE/ Ie exposed individuals to become diseased relative to unexposed? Attributable risk (risk What is the incidence of disease AR = IE- Ie difference) attributable to exposure? Population attributable What is the incidence of disease in Arp = AR x P risk a population associated with the occurrence of a risk factor? Population attributable What fraction of disease in a AFp = Arp/RT fraction population is attributable to exposure to a risk factor?  Relative Risk  Also known as Risk Ratio  RR is the standard measure of association for cohort studies  RR is an index of strength of the association between a risk factor and disease (causal relationship)  RR = Incidence in exposed (IE) Incidence in non-exposed (Ie)  Using the 2x2 contingency table RR = a / (a + b) c / (c + d) If no additional risk is associated with exposure, then both incidences should be equal and the ratio is equal to one (1)  Disease + - Total Exp + a b a+b IE = a/ a+b Exp - c d c+d Ie= c/ c+d Data from a fixed cohort study of oral contraceptive use (OC) and myocardial infarction (MI) in pre-menopausal women followed for 5 years MI (+) (-) Total Yes 23 304 327 OC use No 133 2816 2949 Total 156 3120 3276  RR - Interpretation RR = 1.0 – indicates the rate (risk) of disease among exposed and non-exposed are identical – No association RR = > 1.0 – Risk is higher among exposed – Positive association – Larger RR- the stronger the association RR = < 1.0 – Risk is lower in exposed – Negative Association (factor is protective)  Risk of Gastroenteritis By consumption of a particular food item (Chicken) Ill Not Ill Total Incidence Ate Chicken Yes 43 11 54 80% No 3 18 21 14% 56 29 75 Risk ratio = 80 / 14 = 5.7  Relative Risk Number of people Number of people who who ate specified item did not eat specified item Incid Attack Incid Relative Attack Food Ill Well Total ence% Ill Well Total ence% Rate Rate Risk Baked Ham 29 17 46 63 17 12 29 59 1.07 Mashed potatoes 23 14 37 62 23 14 37 62 1.00 Spinach 26 17 43 60 20 12 32 62 0.97 Cabbage Salad 18 10 28 64 28 19 47 60 1.07 Milk 2 2 4 50 44 27 71 62 0.81 chicken 43 11 54 80 3 18 21 14 5.71 Fruit salad 4 2 6 67 42 27 69 61 1.10 Ice Cream (Choc) 25 22 47 53 20 7 27 74 0.72 Source: CDC Excite  Attributable risk (AR)/ risk difference (RD) AR = IE – Ie Attributable risk is the additional incidence of disease attributable to the risk/exposure itself AR = Incidence in population if all cases are associated with risk factor AR = 0 when there is no association between exposure and disease  RR vs RD Relative risk is a measure of strength of association between exposure and disease and is useful in analytical studies – risk Risk difference is a measure of how much disease incidence is attributable to exposure – burden  Population attributable risk (ARp) Arp = AR x prevalence of risk factor in the population(P) useful in deciding which factors are important and which are trivial in the overall incidence of particular disease in the herd It provides a measure of how much a risk factor contributes to the disease incidence at the pop’n A relatively weak risk factor that is quite prevalent could contribute more to disease incidence in a population than a stronger risk factor that is rarely present.  Population Attributable Fraction (AFp) Population Attributable Fraction (AFp) AFp = Arp / Total incidence of disease in a population (RT) It permits to predict the proportion of cases of a particular disease that will be eliminated through control of a particular risk factor If all cases are associated with the risk factor being measured, then the AFp = 1.00 or 100% Cohort study of risk in neonatal calves with different levels of serum gamma globulin Gamma Cohort size Deaths/ Incidence (%) globulin (%) culls 1.1 – 6.2* 73 12 16.44 6.3- 46.7 220 6 2.73 Totals 293 18 6.14 *Low gamma globulin  Calculation of the measures of effect Simple risk: 1. Incidence of calf losses among low gamma globulin group (exposed/with risk factor) 2. Incidence of calf losses among remaining calves 3. Prevalence of low gamma globulin levels in all calves 4. Incidence of calf losses  Compared Risks: 1. Relative risk = 16.44 / 2.73 = 6.03 2. Attributable risk = 16.44 – 2.73 = 13.71% 3. Population attributable risk = 0.137 x 0.249 = 0.0341 or 3.41% 4. Population attributable fraction = 0.0341/0.0614 = 0.556 0r 55.6%  Interpretation Calves with low serum gamma globulin levels are approximately six times more likely to be culled or die relative to their “normal” counterparts (relative risk) Low serum gamma globulin levels are associated with an additional 13.71% incidence of culls and deaths among exposed calves (attributed risk) Low serum gamma globulin levels are associated with an additional 3.41% incidence of culls and deaths among all calves (population attributable risk) Low serum gamma globulin levels are associated with approximately 56% of all calf losses among all calves (Population attributable fraction)  Comparing risk in case-control studies Since case control study begins with the selection of cases, there is no data on the size of the population at risk, and consequently incidence of disease Participants are selected for study participation on the basis of pre-existing disease status Cannot estimate prevalence, CI – Do not know the population at risk It is not possible to obtain relative risk, we can only make an estimate – ODDS RATIO – ratio of odds (likelihood;chance) of exposure among cases to odds of exposure among controls  Odds Ratio Defined as the ratio of the (number of diseased animals exposed to a factor multiplied by the number of non-diseased animals not exposed to the factor) to (number of non-diseased animals exposed to the factor multiplied by the number of diseased animals not exposed to the factor) OR = (a/c)/(b/d) = ad/bc  Odds ratio Cases Non- cases Exposed A B A +B Not exposed C D C +D A+C B+D Calculation of odds ratio using data from the previous table as if the study were a case control study Cases Non-cases Total Expose to low 12 61 73 gamma globulin levels Not exposed to 6 214 220 low gamma globulin levels Total 18 275 293 ODDS RATIO = AD/BC =(12x214) /(61x6) = 7.02  Summary Risk factors are characteristics that are associated with the increased likelihood of an event occurring (becoming diseased). Whether or not the risk factor is a cause of disease, its presence allows one to predict the probability that the disease will occur. Many suspected risk cannot be manipulated for the purpose of an experiment, so it is usually necessary to study risk by simply observing individuals experience with the risk factor. One way to do it is to select cohort and observe the subsequent incidence of disease. When disease rates are compared among groups with different exposures to a risk factor, the results can be expressed in several ways: RR, AR, Arp, AFp.