Popular in Fundamentals of Epidemiology
Popular in Public Health
This 14 page Bundle was uploaded by Alisha Monnette on Sunday February 1, 2015. The Bundle belongs to EPH521 at University of Texas at Austin taught by Dr. Hlaing in Fall. Since its upload, it has received 118 views. For similar materials see Fundamentals of Epidemiology in Public Health at University of Texas at Austin.
Reviews for Epi_Fall14
These were really helpful...I'll be checking back regularly for these
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 02/01/15
Fundamentals of Epidemiology 02012015 Epidemiology Foundational science of public health and preventative medicine Study of the distribution and determinants of disease frequency in populations Application of study is to control health problems Preliminary association is what we call prevention 0 Frequency 0 Frequency is the occurrence 0 Simple count difference between risk and rate 0 Prevalence rate is a proportion 0 Pattern of disease 0 Person time and place 0 Distribution 0 Determinants 0 Causes 0 Can be risk or preventive Human Population Outcome does not have to be disease it can be a health phenomenon o Morality 0 Quality of life 0 Maternal child health issues Uses of Epi 8 factors in notes 7 factors in book Figure out which is missing and why What is the 2 main purpose of Epi 0 Need to know the cause 0 Do Etiologic research do determine the cause Risk and prevention factors 0 Need to know what is going on in the community Monitoring the trend H1N1 Epi in the past was focused on infectious disease 8 3 levels of prevention based on natural history of disease 0 active primary prevention wearing a seatbelt o prepathogenesis o passive primary prevention uoride in the water so you don t have to do it on your own secondary prevention health screening 0 coonoscopy to prevent colon cancer 0 most cancer screening are secondary pathogenesis a Person has family history of cancer Go to the dr for screening it is primary Tertiary prevention prevent further damage and or restore someone s quality of life rehab halfway house Exposure variable independent variable Condom use Outcome variable disease variable dependent variable 0 STD Target population Ex All undergrad college students 0 Source population Ex From two college campuses undergrad 0 Study sample sample size Ex 100 college students 0 Always include target source and study in research question Association does not mean causation Before we get to this conclusion we have to do many studies to determine if ex Smoking causes lung cancer ABSTRACT Exposure cigarette smoking 0 Outcome CHD both fatal and non fatal 0 Study Sample 1394 men between ages 65 and 74 0 Source Population men in Honolulu heart program 0 Target population elderly men 0 Target population is the one you would like to make the inference about 0 Research question 0 Statement 0 Question 0 Null hypothesis Epidemic and outbreak are synonyms 13 Ebola virus is an epidemic In excess of the expected rate Endemic is what is usually expected can change depending on areas What might be endemic in one area might be epidemic in another area Sometimes we may have the same actuality of rates but the conditions for each rate are different Rate is always person time at risk Prevalence has new cases and previously existing cases Determined by incidence and duration of disease 0 P Incidence x Duration Incidences has only new cases quotAging inquot age group of over 50 yrs old with time you age into certain groups Prevalence Proportion 01 Existing new burden Cumulative Incidence PROPORTION Incidence Density RATE Prevalence is a function of both new cases and the avg duration of disease Pancreatic cancer incidence is very low rapid disease so duration is very short thus prevalence is very low close to incidence Hype engon 0 Incidence high 0 Duration super high Cumulative incidence is a proportion Used in a xed population n remains the same throughout study 0 Numerator of new cases 0 Denominator Population at risk PAR people who are susceptible to getting the disease Incidence density is a rate 0 Used in a dynamic population people come and go o Numerator of new cases Crude a crude death of the US crude death between Miami Dade county and broward Don t give any attention to demographic Speci c what is death of women in a given population Years of Potential Life Lost YPPL always compare to life expectancy If it s negative you count that as 0 DisabilityAdjusted Life Years you do not need to know how to calculate it includes people who suffer from disease but did not die Includes life living with disability Use indirect method when you don t have complete information or when the size of your study is too small thus it is unstable Week 5 Crude rate is used in a more broader sense Use indirect when you don t have a speci c age population you are using Primary data you collect it is expensive labor intensive but you control what you wanna collect and how you collect and know if its reliable Secondary data someone else collects but it may not be the exact answer to your research question Census is important for epi because it gives us our demoninator data You need population size very useful data source Vital statistics data that is collected by National Vital Statistics System CDC Birth death marriage and divorce Behavior risk surveillance Is for adults Youth risk is for young people Epi goal to monitor trends of disease or risk factors past present and future Surveillance is legally mandated Registry is more expensive some funded by federal agencies or private foundations Surveillance does everything research does only thing different is that it is ongoing Descriptive epi person or what place or time characteristics 0 Monitoring trends Analytics epi how and why Hispanic paradox NonHispanic whites have more disease but they die less Strengths of census 0 general measure of population size gives us denominator population midyear population Homeless immigrant elderly and prison populations are not included Limitation of death certificate cause of death may not be accurate due to history of patient someone dies with multiple diseases Difference between surveillance and research where Surveillance is ongoing research you do for a certain period of time o Surveillance active go out and get info 0 Surveillance passive most ppl practice this it is cheaper retrieve info from government Registry optional enrollment that is ongoing Cancer registry registry follows ppl throughout disease very detailed and expensive Cancer surveillance collect ppl who are diagnosed Know HIERARCHY of Study Types Descriptive studying population by person place or time who when Case Report one patient very interesting or very serious Case series more than one case not epidemiological study because there is no comparison Ecological main limitation 0 Geographic or spatial o Advantage lowcost cover large geographic area 0 Sometimes called correlational study Crosssectional analysis is individual temporary relationship cannot be determined because we measure all at the same time Analytic how and why Observational 0 Cross sectional o Casecontrol good for rare disease and all diseases that seek medical care 0 Cohort studies 0 Experimental do manipulation and randomization o Randomized controlled trials manipulate exposure 0 Random selection giving people the same chance ex Put names in a hat equal chance 0 Randomization comes after random selection randomize selected participants in study 0 Quasiexperimental only manipulation no randomization Probability sampling 0 Simple random sampling is when you give everyone an equal random chance 0 Systematic sampling based on a rule that you specify before you start sampling Strati ed sample you can make your group into different groups depending on what factors you think are important Cluster sampling Nonprobability sampling 0 We don t know the sampling population we don t know probability it is not random its convenient Know basic design most commonly used When exposure variable is constant ovrtik Incidence is new cases prevalence is existing Incidence is better because it can help you nd the exposure cases Controls pick controls from which cases arrive not disease free just free of disease that you are interested in studying Pick controls from the same place you are getting the cases 2 stages of study design bias Selection bias how we select the study sample into our study Information bias OR gt 1 is a risk factor OR lt 1 protective factor OR 1 no association Concordited pairs they do not contribute to OR calculations Discorted pairs Cohort study is an observational study because it doesn t have manipulation It s a analytic study because you are testing an hypothesis Two types of cohorts Closed cohort xed cohort xed membership initial roster may dwindle as ppl die 0 Open cohort dynamic cohort can take on new members as time passes Three different types of cohort studies Prospective cohort start at baseline 0 and observe prospectively in real time Retrospective cohort looking back on cases through historical time to present although you look back both disease and exposure started before you re experiment study begin Ambispective cohort both prospective and retrospective part You cannot estimate prevalence or incidence from casecontrol studies You cant estimate prevalence from case control studies because we determine how many casescontrols we take Crosssectional you can estimate prevalence because you select randomly Biggest limitation of cohort it takes a long time expensive live sample longer the study more likely the study population will dropout you cannot study rare disease Prospective Cohort Retrospective 1 Start with disease free sample 2 Exposure 3 Followup 4 Determine outcome Selection bias begins with selecting ppl for your study Information bias already selected ppl and now you have to collect data from them Interviewer bias from ppl who collect the data 0 Recall bias from subjects themselves We select population from exposure status exposednon exposed and follow them up Week 8 cohort NOT on midterm Look at detailed course schedule Read chpts 4 amp 5 0 Descriptive epi 0 Infant Mortality Ratio Exam sample questions Experimental manipulation and randomization Quasiexperimental is only manipulation Clinical Trials Random selection selecting ppl random for clinical trial 0 Random allocationassignment Control Double blind Experimental Allocation Historical Controls using patients from previous treatments Systematic Assignment on the basis of some rulefactor Random allocation ooqoo Clinical trials have to be at least double blind Once you randomize patients you must analyze They should be included regardless of compliance 0 Compliance issue is big part of clinical trials issue 0 Extent to which your patient listens to you it undermines the bene t of treatment People don t comply because of side effects it s complicated Pay attention to DSMB and consort statement pg 390 you don t need to know xed and adaptive randomization you don t need to know how to evaluate community trials nor cross over study design Pg 390 NIH requires oversight and monitoring of all intervention studies to ensure the safety of participants and the validity and integrity of data This goal is accomplished through the DSMB Data and Safety Monitoring Board DSMB is composed of an independent group of experts who advise investigators and funding agencies Primary responsibilities of DMSB are Periodically review and evaluate the accumulated study data for participant safety study conduct and progress and when appropriate ef cacy 0 Make recommendations to Division of Microbiology and Infectious Disease DMID concerning the continuation modi cation or termination of the trial 0 Although phase III clinical trials are required to have a DSMB phase I amp II trials in certain circumstances may also require them Results of RCT randomized controlled clinical trial can impact a patient greatly Readers need to be able to comprehend key aspects of the trial design conduct and analysis as well as to determine the general ndings To meet this need a panel of experts developed the CONSORT statement Consolidated Standards of Reporting Trials 0 22item checklist and owchart to be used by those who review write and assess ndings in RTC 0 main aims of CONSORT 0 help authors optimize the quality of their reports of simple RCTs 0 Ex Can describe the ow of subjects through the phases of an RCT REVIEW WEEK 11 Difference between AR and PAR PAR is for the population it is absolute AR is for exposed among the exposed Difference between AR amp AR AR excess number of cases AR excess proportion of cases Alternative Chance or random error Bias cannot be controlled Prevalenceincidence bias survival bias Case control amp cross sectional study Prevalent case only ppl who are still alive Confounding must be associated with exposure and disease 0 It cannot be in the causal pathway between exposure and disease Confounding has to be differentially distributed How do we detect confounding with data 0 Stratify analysis Difference between potential and confounder potential you don t have data yet Bias can be prevented and controlled lnternal validity is more important than external Strength amp Dose response Validity is a measurement of quality Go over brick BEASTSSCDC Screening should be rapid inexpensive o What is the survival Leadtime bias 0 Two ppl die at same time but because one is detected earlier you determine a leadtime bias 0 Length bias VERB campaign it s what you do social marketing Cutoff values affect lnverse relationship between the two 0 Sensitivity Speci ty Positive predictive value Endemic expected rate that is commonly present Epidemic higher than normal excess Pandemic crosses geographic boundaries goes global Outbreak is a synonym for Epidemic Before you calculate Relative risk you need to calculate attack rate What is secondary attack rate If RR 1 there is no need to calculate AR PAR ect Friis has good example for how to calculate attack rate and secondary act rate
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'