Epi Final Spring 2024

• Random Error


  Chance = uncontrollable force with no apparent cause that arises due to unforeseeable and unpredictable processes
• Random Error


  • Can occur in all types of epidemiologic studies
  • Caused by measurement error, sampling variability
  • All studies have some random error
• Sampling Variability


  • Study subjects are a sample from a population of interest
  • Statistical inference: process of making statements about a population based on information from a single sample
• Precision


  Lack of random error; state of being precise or exact
• Approaches to reduce random error and increase precision in an epidemiological study

  • Use an accurate measurement tool; repeat measurements
  • Increase sample size
• Precision is not the same as accuracy
• The absence or reduction of random errors ≠ The absence or reduction of systematic errors
• Hypothesis Testing


  • Specify a null (H0 ) and alternative (HA)hypothesis
  • Use a statistical test to determine the compatibility of study results with null hypothesis
  • Make a decision whether or not to reject the null hypothesis based on the p value associated with the statistical test
• 1. values


  • Probabilities: continuous measure ranging from 0 to 1
  • If there is truly no association between the exposure and disease, then there is a X% probability that we would have obtained the results we did or results that are more extreme
  • P-values are calculated assuming H0 is true

  
• Interpreting P-values


  • Small p-value indicates low degree of compatibility between null hypothesis of no association and observed result, reject the null hypothesis
  • Large p-value indicates high degree of compatibility between null hypothesis of no association and observed result, don't reject the null hypothesis
• Statistical significance ≠ Causal association : P-values tell us about the compatibility with the null, not compatibility with the truth
• What P-values are not


  • The probability that the null hypothesis is true
  • Used to prove that the null hypothesis is true
  • Used to prove that an exposure causes a disease
  • The only number you need to report
  • Used to imply medical, biological, or public health significance
• Limitations of the P-value

  • P-values are confounded (mix together magnitude of association and sample size)
  • P-values do not rule out bias or confounding
  • Over-reliance on 0.05 cut-point leads to errors
• Confidence Interval (CI)


  Range of values within which the true magnitude of association lies with a stated probability (e.g., 95%) given our observed data and assuming no bias or confounding
• Most epidemiologists prefer confidence intervals over p-values
• CIs do a better job of separating magnitude from sample size
• Steps for Critical Review of Epidemiological Studies

  • Collection of data
  • Analysis of data
  • Interpretation of data
• Key principles of ethics in research involving human participants

  • Respect for individual autonomy
  • Beneficence
  • Justice
• Institutional Review Boards (IRBs) currently review research to ensure its ethical conduct
• Informed consent


  Process by which an individual voluntarily expresses his/her willingness to participate in research, after having been informed of all aspects of the research that are relevant to his/her decision
• Key considerations for informed consent


  • Information exchange
  • Comprehension
  • Voluntariness
  • Documentation
• Informed consent is not a form or a signature but a process
• Elements of information exchange in informed consent

  • Statement that the study involves research, an explanation of the purposes, duration, procedures, and identification of experimental procedures
  • Description of any reasonably foreseeable risks or discomforts
  • Description of any benefits
  • Disclosure of appropriate alternative procedures or courses of treatment
  • Statement describing the extent of confidentiality
  • Explanation of compensation and medical treatments available if injury occurs
  • Explanation of whom to contact for questions and in case of research-related injury
  • Statement that participation is voluntary
• It is the responsibility of the investigator to ensure comprehension of informed consent
• It is the responsibility of the investigator to assure that informed consent is obtained in a setting free of coercion and undue influence
• It is the responsibility of the investigator to document the informed consent process
• Three essential attributes of a cause

  • Association: causal factor must occur together with its effect
  • Time Order: cause must precede effect
  • Direction: asymmetrical relationship between cause and effect
• Epidemiology


  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.

  
• Goal of Epidemiology


  • Identify causes and preventions for disease
• Building Blocks of Epidemiology

  • Measures of Disease Frequency
  • Various Study Designs
• Measure of Association


  A statistic that quantifies the relationship between an exposure and an outcome
• Once you have calculated a measure of association, you need to determine if the observed association is valid and if it is causal. This process is called "causal inference."
• Association is not causation
• Three Essential Attributes of a Cause

  • Association: causal factor must occur together with its effect
  • Time Order: cause must precede effect (can be either proximate or distant in time)
  • Direction: asymmetrical relationship between cause and effect
• Causal Web


  Developed in 1960s, many interconnected factors (both host and environmental) for the cause of diseases, represents an important paradigm shift where disease causes are conceptualized as multi-factorial in nature. This model fits better for non-infectious diseases.
• Sufficient Cause Model (SCM)

  Another multi-factorial causal model where "Causes" are conceptualized as pies and pie pieces. All of the pie pieces need to fall into place for disease to occur. The whole pie is called the "sufficient cause." A "sufficient cause" is a complete causal mechanism that inevitably produces disease. The pie pieces are called "component causes" which are participating factors in a sufficient cause and are the focus of much of our research.

  
• Features of SCM


  • Each pie represents a sufficient cause: a complete causal mechanism that results in a particular outcome
  • Each wedge represents an etiologic factor and is called a component cause
  • Component causes may be behavioral, environmental, or genetic
  • Note that component cause "A" is present in all of the sufficient causes above. These are termed necessary causes.

  
• Sir Austin Bradford Hill


  Proposed a set of guidelines for assessing causality, nine items to help guide our "causal inference" but do not provide indisputable evidence for or against causation
• Hill's Guidelines


  • Strength of Association
  • Consistency
  • Specificity
  • Temporality
  • Biological Gradient
  • Plausibility
  • Coherence
  • Experiment
  • Analogy
• Primary prevention


  Preventive measures that prevent the onset of illness or injury before the disease process begins, e.g. immunization and taking regular exercise