keywords:
Bookmark and Share



Front Back
Define rates.
Rates determine the number of actual cases in comparison to the number of potential cases. Rates are generally per 100,000.
Define incidence and prevalence.
Incidence rate is the rate at which new events occur on a population. Attack rate is a type of incidence rate in which the denominator is further reduced for some known exposure.
Prevalence rate is all persons who experience an event in a population. Point prevalence is prevalence at a specified point in time. Period prevalence is prevalence during a specified period.
Morbidity rate is the rate of disease in a population at risk.
Mortality rate is the rate of death in a population at risk.
Describe the relationship between incidence and prevalence.
Prevalence = Incidence × Duration (P = I × D)
New cases are monitored over time. New cases join pre-existing cases to make up total prevalence. Prevalent cases leave the prevalence pot in one of two ways: recovery or death.
Define crude, specific and standardized rates.
Crude rate is the actual measured rate for the whole population.
Specific rate is the actual measured rate for a subgroup of population.
Standardized rate (or adjusted rate) is adjusted to make groups equal on some factor.
Define sensitivity, specificity, positive predictive value, negative predictive value and accuracy.
Sensitivity is the probability of correctly identifying a case of disease. (SNAP: sensitivity identifies positive.) Sensitivity = TP ÷ (TP + FN)
Specificity is the probability of correctly identifying disease-free persons. (SPIN: specifity identifies negative.) Specificity = TN ÷ (TN + FP)
Positive predictive value is the probability of disease in a person who receives a positive test result. PPV = TP ÷ (TP + FP)
Negative predictive value is the probability of no disease in a person who receives a negative test result. NNV = TN ÷ (TN + FN)
Accuracy is the deegree to which a measurement represents the value of the attribute that is being measured. Accuracy = (TP + TN) ÷ (TP + TN + FP + FN)
Selection bias
  • Sample not representative
  • Berksons's bias, nonrespondant bias
  • Random, independent sample
Measurement bias
  • Gathering the information distorts it
  • Hawthorne effect
  • Control group / placebo group
 
Experimenter expectancy bias
  • Researcher's beliefs affect outcome
  • Pygmalion effect
  • Double-blinded design
 
Lead-time bias
  • Early detection confused with increased survival
  • Benefits of screening
  • Measure "back-end survival"
 
Recall bias
  • Subjects cannot remember accurately
  • Retrospective studies
  • Multiple sources to confirm information
 
Late-look bias
  • Severely diseased individuals are not uncovered
  • Early mortality
  • Stratify by severity
 
Confounding bias
  • Unanticipated factors obscure results
  • Hidden factors affect results
  • Multiple studies, good research design
 
Design bias
  • Parts of study do not fit together
  • Non-comparable control group
  • Random asignment
 
Random error is unfortunate but okay and expected, and is a threat to...
Systematic error is bad and biases result, and is a threat to...
...reliability.
...validity.
Cross-sectional studies
  • Time: one time point
  • Incidence: no
  • Prevalence: yes
  • Causality: no
  • Role of disease: prevalence of disease
  • Assesses: association of risk factor and disease
  • Data analysis: chi-square to assess association
 
Case-control studies
  • Time: retrospective
  • Incidence: no
  • Prevalence: no
  • Causality: yes
  • Role of disease: begin with disease
  • Assesses: many risk factors for single disease
  • Data analysis: odds ratio to estimate risk
 
Cohort studies
  • Time: prospective
  • Incidence: yes
  • Prevalence: no
  • Causality: yes
  • Role of disease: end with disease
  • Assesses: single risk factor affecting many diseases
  • Data analysis: relative risk to estimate risk
 
x of y cards