Video summary

8ª AULA - BIOESTATÍSTICA E EPIDEMIOLOGIA - 52/25

Main summary

Key takeaways

Educational

Main ideas & lessons

  • Course context and roadmap

    • This is the 8th lesson in Biostatistics and Epidemiology.
    • Prior units covered:
      • Unit 1–3: descriptive statistics (Ch. 1–3 of the book)
      • Unit 4: inferential statistics (correlation, linear regression, statistical inference)
    • Reading assigned before this class:
      • Unit 5, 6, 9
    • Today’s focus:
      • Unit 6 (main focus) and a bit of Unit 7
    • Next class likely continues Unit 7 and 8.

  • Learning objective for today

    • Study health indicators
    • Calculate them
    • Interpret them correctly (emphasis on interpretation)

  • What epidemiology is

    • Etymology:
      • Epi = upon
      • Demo = population
      • Logos/Logia = study
    • Epidemiology described as:
      • an information science in health
      • foundational for medicine, public health, and other health professions
    • Core practice:
      • collect health information in a systematic/standardized way
      • use defined criteria to ensure comparability across populations and settings

  • Bias and study design (quality requirement)

    • Bias is framed as an error from the researcher’s side, e.g., selecting samples that fit the researcher’s needs rather than representing the population.
    • Epidemiological studies require comprehensive/population-representative data collection to avoid bias.

  • Choosing health indicators

    • Requires prior understanding of:
      • the population and/or disease
    • Indicators must be selected according to defined methodological and ethical considerations:
      • there is usually a regulatory body defining indicators and definitions
      • ethics: some patient health data may be confidential and cannot be disclosed
      • operational feasibility: some indicators may be easier/harder to apply in practice

  • Reliability and reproducibility

    • Epidemiological indicators must be:
      • highly reliable
      • highly reproducible
    • Reproducibility means:
      • if the study is repeated in another region under the same criteria, results should be comparable.

Methodology / workflow (data → analysis → interpretation)

  • Overall approach
    • Observation of the health-related phenomenon
    • Data collection using standardized criteria
    • Coding/processing of collected data
    • Analysis/outputs, including:
      • tables and graphs
      • descriptive metrics (mean, standard deviation, variance)
      • inferential tools (confidence intervals, hypothesis testing)
    • Interpretation stage
      • explain what the calculated indicators imply for the studied population

Key health indicators and what they mean

Mortality indicators

  • Mortality rate concept
    • Measures the risk of death in a population/region over a period.
    • Used to:
      • track disease trends over time
      • compare regions’ health levels
      • detect possible low quality in care (e.g., prenatal/childbirth/postpartum care)
      • support public investment decisions (e.g., sanitation, public health)
  • Common unit expressions
    • Expressed typically per 1,000 or per 100,000 inhabitants per year (and related scaling multiples may appear).
  • Types of mortality
    • Specific mortality rates
      • e.g., pancreatic cancer mortality; mortality due to coronary heart disease
  • Case fatality rate (CFR)
    • Deaths caused by a specific disease / total confirmed cases of that disease
    • COVID-19 example: deaths from COVID-19 divided by total confirmed COVID-19 cases
  • Proportional mortality rate
    • Deaths from a specific cause / total deaths
  • Infant mortality rate
    • Deaths of children under 1 year / number of live births
    • Widely used in Brazil

Morbidity indicators

  • Morbidity
    • A rate/ratio for the number of people who acquire a disease in a given time interval within a population.
    • Interpreted as:
      • risk of becoming ill
      • helps identify determinants and guide prevention actions
    • COVID-19 example:
      • used to decide what prevention actions management teams should take

Detailed list of exercises / conceptual checks (as presented)

Exercise 1 (theoretical): “Mark the incorrect option” — Epidemiology objectives

  • A (Correct)
    • Describes the distribution and magnitude of health problems in human populations.
  • B (Correct)
    • Studies factors that determine the frequency and distribution of diseases in human communities.
  • C (Incorrect)
    • Claims epidemiology provides data for planning prevention/control but not for treatment.
    • Instructor argument: epidemiology information can support treatment decisions and evaluation of medication effectiveness (studies/hypotheses to find medicines/drugs).

Exercise 2 (theoretical): “Mark the incorrect alternative” — Health indicators in epidemiology

  • Key rule emphasized
    • Health indicators are defined/regulated by some regulatory body (examples like WHO, Ministry of Health, Anvisa mentioned later).
  • Reliability/reproducibility requirement
    • Indicators should yield reproducible and comparable results across regions under consistent criteria.

Note: The transcript doesn’t clearly finish the full multi-letter “mark incorrect option” set at this point; it mainly explains what indicators must satisfy.

Exercise 3: Crude vs age-adjusted mortality (coronary heart disease)

  • Given
    • Crude mortality rate for both municipality A and B: 4 per 1000
    • Age-adjusted mortality:
      • municipality A: 5 per 1000
      • municipality B: 3 per 1000
  • Conclusion logic emphasized
    • Age adjustment reveals differences in population structure (young vs adult proportions).
    • If the adjusted rate rises for municipality A (4 → 5), instructor interprets it as indicating:
      • A has a relatively younger population than B,
      • based on ratio/division reasoning: an increasing division result implies a relative change in the denominator (more of the lower-risk group vs fewer of the higher-risk group).

Exercise 4: Infant mortality and IBGE data

  • Reinforced definition
    • Infant mortality = deaths of children under 1 year / live births (scaled per 1,000 or other multiples).

Exercise 5: Maternal mortality — “Mark the incorrect statement”

  • A (Correct)
    • Uses live birth information system as a source.
  • B (Correct)
    • Considers deaths related to pregnancy, childbirth, and postpartum.
  • C (Incorrect)
    • Claims deaths up to 10 days after end of pregnancy.
    • Instructor correction:
      • the time window is different (transcript suggests a later postpartum timeframe, but the exact wording is not fully consistent).

Exercise 6: Crude birth rate — “Mark the incorrect alternative”

  • B (Correct)
    • Expresses intensity of birth rate in a population.
  • A (Correct concept, but treated as “incorrect” in the transcript due to option-matching confusion)
    • Instructor corrected an option about influence of population structure/age/sex, stating that influence is true.
  • C (Incorrect)
    • Claims low rates necessarily relate to poor socioeconomic conditions and cultural aspects.
    • Instructor: not necessarily true—it depends on context/study design and bias control.
  • D (Correct)
    • Can compute natural growth by subtracting crude mortality from crude birth rate.

Note: The transcript includes internal “wrong/correct” toggling while the instructor discusses options; the takeaway is which conceptual statement the instructor considered incorrect.

Measure of frequency and association (final section)

  • Measures of frequency/occurrence

    • Gauge occurrence of an event (e.g., death, illness, birth) in a population:
      • by group
      • at a place
      • at a time period
    • Event can be expressed as:
      • rate, proportion, percentage, or quantity

  • Observational studies (no intervention by researcher)

    • Researcher only observes natural development of the event.
    • Example: retinal degenerative disease monitored without intervention until the situation changes (study type could change afterward).

Prevalence

  • Definition (as stated)
    • number of people affected by a health phenomenon at a given time
    • / total population at that time
  • Formula form given
    • (sick people including old + new cases) / population × 1000
    • (scaling mentioned; ×100 for percentage is also stated)
  • Interpretation
    • Retrospective concept: includes old and new cases when measured.

Incidence

  • Broad conversational/Q&A definition
    • number of new cases / population at risk
  • Instructor note: incidence will be discussed more in the next class.

Factors affecting prevalence direction (as stated)

  • Prevalence increases when
    • disease lasts longer
    • survival increases (even without cure)
    • more cases enter (immigration of cases/susceptible people)
    • improved diagnostic resources
  • Prevalence decreases when
    • illness is short-lived
    • higher fatality rate
    • fewer cases (reduction), including emigration of healthy people or cases
    • improved cure (vaccines/medications)

Exercise 7/8 (malaria prevalence) — interpret a number

  • Given
    • 6,396 new malaria cases in the first two months of 2005 (Manaus / Tropical Medicine Institute)
  • Instructor’s point
    • That number corresponds to the numerator of the prevalence rate in the specific interpretation exercise (i.e., the “cases” portion).

Homework exercises

  • Exercise 9 (homework)
    • Theoretical prevalence-rate definition-based problem.
  • Exercise 10 (homework, demonstrated)
    • Hypothetical municipality:
      • 2008–2022: 2,450 diabetes mellitus cases
      • population: 250,000 inhabitants
    • Calculation
      • prevalence = 2,450 / 250,000 = 0.0098
      • scaled:
        • × 1,000 → 9.8 per 1,000
        • expressed as whole number using a “multiple-of-10” approach
        • equivalently 980 per 100,000 mentioned (any multiple-of-10 scaling is acceptable)
  • Exercise 11 (homework, more complex)
    • Compute prevalence for:
      • current smokers
      • non-smokers
      • former smokers
    • Then compute prevalence ratio and select the correct alternative.

Q&A points from the chat (notable concepts)

  • Sampling in epidemiology
    • Yes, samples are used in epidemiology.
  • Effect of missing/incorrect indicator data
    • Can compromise results; careful planning and correct indicator methodology are essential.
  • Regulatory bodies
    • Examples: WHO, Anvisa, Ministry of Health (and other institutions may define standards).
  • Do participants need to know the indicators?
    • Typically no; participants join studies and institutions collect/store the data.
  • Prevalence vs incidence
    • Can be confusing: COVID-19 “cases” could be prevalence or incidence depending on data and time structure.
  • Scaling of prevalence
    • Generally by a multiple of 10; health field commonly uses whole-number scaled rates (e.g., per 1,000 or 100,000), though percentages (×100) may appear.

Speakers / sources featured (as named in the transcript)

Speakers

  • Professor Ricardo (main instructor)
  • Adriana (appears as someone greeted; likely a participant/assistant)
  • Amanda (asked a question during Q&A)

Sources / organizations mentioned

  • WHO (World Health Organization)
  • Anvisa
  • Ministry of Health (Brazil)
  • IBGE (Brazilian Institute of Geography and Statistics)
  • Agência Brasil
  • Rede Globo (used as an analogy for retrospective broadcasting)
  • Institute of Tropical Medicine in Manaus
  • COVID-19 pandemic (context reference)
  • Ministry of Health information systems (including the live births system mentioned)

Original video