epidemiology

• Why do we need research?

  • poor quality research leads to poor health outcomes
  • where as quality research leads to healthier populations
• Why don't we invest more on research?

  • cost (low income countries lack funds for research)
  • short-term thinking
  • lack of value recognition
  • technical skills gap
  • pressure for quick results
  • fragmented priorities (investments become scattered)
• What is done in epidemiological research

  • interested in understanding correlations, associations and casual relationships between exposure and outcome
• Exposure

  • a risk factor of interest in studying for potential effects on a certain disease/ other outcomes
  • typically a characteristic, behaviour, condition or treatment individuals are exposed to
• Outcome:

  • Can be a specific disease, condition, or measure that we are interested in studying to examine the impact of the exposure.
  • This can include any health/well-being conditions e.g. quality of life measures, and mortality rates.
• Experimental study design:

  • RCTs (lab trials/ field trials)
  • non-RCTs
• Observational study design:

  • longitudinal
  • case-control
  • cross-sectional
  • ecological
• Longitudinal study:

  • follows individuals over time to examine changes in variables of interest and investigate relationships between them
  • involves collecting data from the same individuals over different time points
• Case-control study:

  • compares individuals with a specific outcome (case) to those without the outcome (controls)
  • used when diseases are spreading quickly and move backward in having the outcome and trying to find the exposure
  • suitable for rare outcomes or diseases
  • prone to recall bias
• Cross-sectional study:

  • most common study in epi
  • collects data from participants at a single point in time to understand the relationship between different variables
  • gives a snapshot of a population's characteristics
  • quick and cost-effective
  • not in-depth and cannot establish causality
• Ecological study:

  • focuses on populations or communities rather than individual-level data
  • findings may not apply to individuals within the population
  • An ecological study analyzes group-level data to explore relationships between population-level variables
  • dedicated to investigating associations within entire populations
• Ecological fallacy:

  • generalize findings from group to individual level
  • ex. higher income leads to better health might not be applicable at the individual level
• Simpson's paradox:

  • extreme condition of confounding in which an apparent association between two variables is reversed when data is analyzed within a confounding table
• Randomized control trials:

  • gold standard for generating valid evidence
  • key features:
  • randomization
  • control group
  • reproducible
  • expensive and various ethical issues
• Prevalence:

  • proportion of individuals with a specific disease/condition in a population at a given point in time
  • indicates the burden/volume of diseased
• Incidence:

  • rate at which new cases of a disease/condition occur in a population over a specific time
  • indicates the the spread of disease spread
• Case-Fatality Rate:

  • proportion of individuals diagnosed with a disease/condition dying because of that
• Mortality rate:

  • number of deaths from a specific cause within a population over a specific period of time
• Measures of association:

  • risk: number of people with the outcome/total number of people at risk
  • odds: number of people with the outcome/ number of people without the outcome
• Risk ratio:

  • compares risk of an outcome (depression) between two groups (alcohol vs no-alcohol)
  • more suited for cohort studies
  • Risk ratio= risk in exposed group/ risk in unexposed group
• Odds ratio:

  • compares the odds of an outcome (depression) between two groups (alcohol vs non-alcohol)
  • more suited for case-control studies
  • Odds Ratio = Odds of outcome in exposed group / Odds of outcome in unexposed group
  • shows correlation not causation
• Interpretation of OR:

  • OR > 1 indicates that the odds of the outcome occurring are higher in the presence of the exposure
  • OR < 1 it indicates that the odds of the outcome occurring are lower in the presence of the exposure
  • OR=1 it indicates the odds of the outcome occurring are the same in the presence of the exposure variable compared to the absence
• Hazard ratio:

  • compares the hazard/risk of an event occurring over time between two groups.
  • Probability of an event happening at a particular time, given it has not happened before
  • suitable for longitudinal studies
  • looking at the outcome but also time it takes for the outcome to happen
• P-value:

  • expression of the probability that the difference between the observed value and the null value has occurred by "chance"
  • the smaller the p-value (the better), the less likely the probability that sampling variability accounts for the difference
• Drivers of food insecurity:

  • individual (poverty, lack of education, disability, unemployment)
  • household (possible to have families where some family members are food secure and others are food insecure)
  • national (poor agriculture output, energy and water insecurity)
  • global (climate change, changing diets)