Study Case Study/Series

Advantages 1. Good for atypical/unusual findings

-presentation of 1 or a series of patients with the same condition or Dx Cross-Sectional Study (Prevalence)

Disadvantages 1. No control or comparison group 2. Cant test for statistical association 1. No temporal relationshipno claims on causation

Other Notes Descriptive Study

-an assessment of the exposure and outcome in an individual at a moment in time Case-Control Study (Retrospective)

1. Quick, inexpensive 2. Conducted among the general populationextern al validity

Descriptive Study

-Most frequently done analytical study -DiseaseExposure

1. Quick, inexpensive 2. Most efficient for rare diseases 3. No loss to followup 4. Can evaluate multiple exposures 5. Can use secondary data

1. Temporal relationship uncertain 2. Cant provide direct estimate of risk 3. Not efficient for rare exposures 4. Subject to recall limitations and biases

Analytical-Observational Study OR (odds ratio or relative odds) =odds of exposure (cases)odds of exposure (controls)

Use CI Bivariate or multivariate analysis (logistic regression)

Interpretation: Those with the disease are OR times more/less likely to have been exposed than not-exposed. ----------------------------------------------Matching: increase the similarity of cases and controls except for exposure of interest Multiple Controls: Same type-increases power (internal validity) Different type-increase external validity Analytical-Observational Study RR (relative risk) =incidence risk disease exposedincidence risk disease non-exposed

Cohort Study (Prospective,Longitudi nal, Incidence, Followup)

***1. Temporal relationship wellestablished 2. Calculate/compare risk (incidence

1. Costly, lengthyrequires large sample 2. Not efficient for rare disease 3. Loss to follow-up

-Most rigorous observational study -Starts with people free of disease BUT susceptible -Exposure Disease Types based on temporal relationships: 1. Prospective 2. Historical Cohort 3. Historical Prospective (Ambicohort)

rates) 3. Most efficient for rare exposures 4. Minimizes assessment/inform ation biases 5. Can evaluate multiple outcomes

AR (attributable risk) =incidence risk diseased (exposed) incidence risk disease (non-exposed)

Interpretation: Those exposed are RR times more/less likely to develop disease than those who were not exposed. ---------------------------------------------Matching less critical compared to case-control

Types of Controls: 1. Internal (single-sample design) 2. External (multi-sample design) 3. Population when a nonexposed comparison group cannot be assembled.

Epidemiology: the study of the distribution (who, what, where when) and determinants (why) of disease in a population Aims: Describe (obtain relative frequencies) Explain (etiology, transmission) Predict Control (prevent, eradicate, prolong lives)

The Hierarchy of Strength of Evidence: Analytical -determinants Experimental -intervention (prevent and treat) Observational -determinants Experimental (Clinical Trial, Field Trial, Community Trial) Cohort Synthesis Systematic Review

Case-Control Descriptive -distribution -tend to raise Qs Case-report/Series Ecological (population-based) Cross-Sectional

Study Variables: 1. Independent variable (risk factor/exposure) 2. Dependent variable (disease/outcome/result) 3. Covariates Confounding variables Effect modifiers

Measures of Disease Frequency: 1. Incidence Rate (per 1000) = NEW cases or events in population in specified time periodNumber at risk in specified time period X 1000 2. Prevalence Rate (per 1000) = Number EXISTING cases or events at specified timeNumber at risk in specified time X 1000 Point prevalence: At a certain movement in time Period prevalence: During a period of time

Types of Rates: 1. Crude: the number of events in a population over a given period of time Actual summary rate Reflects likelihood of event Useful for health services planning and resource allocation (estimation of economic burden on a community)

2. Specific: stratifies on person, place, time and/or cause to eliminate systematic differences in populations being compared To compare rates among subpopulations or for various causes Some are used for public health planning purposes or metrics for overall health of a population

3. Adjusted: allows two populations to be compared as if the two populations were equivalent on the confounding variables Fictional summary rate

To compare the health of an entire population with different demographic structures

Endemic, Epidemic, Pandemic: Endemic: Usual number of cases expected in a given geographic region Epidemic: Unusually high number of cases in a given geographic region Pandemic: Geographically widespread epidemic

Statistics: the discipline which is concerned with data Collection Organization Display Interpretation

Definitions: Population: all of the subjects or observations that belong to the category or class of interest. Investigator defined. Sample: a subset of the population. Must be representative of the population if the results are to be generalizable.

Analysis Strategies: 1. Descriptive Statistics: to summarize and describe the data from a sample without trying to generalize the results to a larger group. Central Tendency: Mean, median, mode Dispersion: Range Variance: the average amount by which individual values deviate from the mean Standard deviation: square root of variance, recovers proper unit of measurement data Vs. Standard Error: measure of precision of the sample statistic (e.g. the mean) Note: Normal Curve: 95% within 1.96 SD or SE from the mean

2. Inferential Statistics: methods and tests used to draw conclusions from the sample data and to generalize the results to the larger population. Hypothesis Testing Run Test: Statistical test depends on: measurement scale, data distribution, sample size Calculate P-value: the likelihood or probability of getting the results observed by chance alone, assuming null hypothesis is true Compare with

Basic structure: Signal=effect size:difference between observed data and what expected under null hypothesisNoise:the variability of the estimate increase in sample size decrease noise Error, etc. Type I Error (): probability we reject null when null is true; usually . 05 Type II Error (): probability we fail to reject null when null is false; usually .20 Power (1-): ability/probability to detect a difference when one exists; usually .80 Depends on the sample size and effect size underpowered if small effect size or too few subjects Sample Size Specify and Magnitude of desired effect (whats clinically useful?) Measurement variability Value in the control or baseline group Confidence Intervals: (usually 95% used) Statistic or point estimate +/- (critical value)(SE of the statistic) Margin of error = (critical value)(SE of statistic)

General Principles for Data-Based Studies: 1. Define the study population 2. Define the features of representative sample Inclusion criteria Exclusion criteria: provides greater internal validity BUT decreases external validity

3. Obtain the sample Convenience sampling Probability sample Simple random Stratified random Systematic random Clustered sampling

4. Data Collection-Measurement, Observer, Subject Precision (reliability, reproducibility)-variabilities Accuracy (validity)-Biases Types: Internal validity External validity