Wykład 2 Advanced biostatistical tools

Advanced Biostatistical Tools

  • Small-area methods: Tools for studying epidemiological risk

  • Importance of assessing the need for these methods in current health studies.

COVID-19 Dashboard

  • Global Cases: 48,092,115

  • Cases by Country/Region/Sovereignty:

    • US: 9,486,486

    • India: 8,364,086

    • Brazil: 5,590,025

    • Russia: 1,680,579

    • France: 1,591,152

    • Spain: 1,284,408

  • Global Deaths: 1,225,221

    • US: 233,729 deaths

    • Brazil: 161,106 deaths

    • India: 124,315 deaths

    • Mexico: 93,228 deaths

    • UK: 47,832 deaths

    • Italy: 39,764 deaths

    • Other countries listed with significant death tolls.

  • US State Level Data:

    • New York: 33,556 deaths, 80,109 recovered

    • Texas: 18,771 deaths (recovered numbers missing)

    • California: 17,820 deaths

    • Florida: 16,922 deaths, 37,559 recovered

    • New Jersey: 16,391 deaths (recovered numbers missing)

  • Dashboard Includes:

    • Active Cases

    • Incidence Rate

    • Case-Fatality Ratio

    • Testing Rate

    • Data sourced from various databases including JHU CSSE.

Small-Area Studies

  • A powerful approach for studying disease patterns in populations.

  • Evaluates health risks posed by environmental pollutants.

  • Includes:

    • Environmental exposures.

    • Health impacts on populations.

Framework of Small-Area Studies

  • Public health investigations are conducted on geographical scales (e.g., neighborhoods).

  • Overlay of health, environmental, demographic, and confounder data is essential.

  • Key aspects include:

    • Rapid data linkage from diverse sources for improved analysis.

    • Utilization in surveillance and prevention of non-communicable diseases.

Methodological Advances in Small-Area Studies

  • Incorporation of Bayesian approaches.

  • Enhanced computational capabilities facilitating more informative analyses.

Complexity and Challenges

  • Increased complexity in design includes:

    • Choice of optimal study area size.

    • Duration of the study.

    • Range of covariates/confounders considered.

    • Managing uncertainty in data.

Requirement of Data Availability

  • Small-area studies rely heavily on the availability of:

    • Health data

    • Environmental data

    • Demographic data

    • Confounder data

  • Focuses on entire populations or large subsets for validity.

Example of Health Data Utilization

  • Neighbourhood data example: deprivation and environmental exposures.

  • Longitudinal studies linking participants to residential address and health data.

Benefits of Small-Area Studies

  • Factors to consider:

    1. Approximate individual-level risks when individual data is scarce.

    2. Investigate health risks from environmental pollution.

    3. Detect high-risk areas for interventions.

    4. Conduct investigations of disease clusters.

Health Data Accessibility

  • Access to health records has greatly improved over the last 2-3 decades.

    • Example from England:

    • 100 million records added annually in NHS.

  • The Small Area Health Statistics Unit (SAHSU) maintains a comprehensive database with over 600 million health records.

Geographical Data Analysis

  • Caveat: Population data may show disparities, especially in rural low/middle-income countries.

  • Importance of population-based cancer registries as data sources.

Addressing Confounder Data

  • Small-area studies require attention to confounding factors to validate findings.

  • Must standardize data across demographics (age and gender).

Spatial Epidemiology Concepts

  • Spatial epidemiology involves examining geographic variations in disease concerning various risk factors:

    • Demographic

    • Environmental

    • Socioeconomic

    • Genetic

    • Behavioral

Disease Mapping

  • Visual representation aids in identifying the geographical distribution of disease.

  • Historical example: John Snow's cholera mapping in London (1854).

Smoothing Techniques in Data Analysis

  • Statistical smoothing aids in reducing variability across geographical data.

  • Allows for more accurate readings but can potential mask risks in less populated areas.

Uncovering Disease Clustering

  • Clustering analysis focuses on identifying unusual case accumulations within a population.

  • Requires distinguishing between general and specific clustering approaches.

Application of Small Area Techniques to COVID-19

  • Small Area Estimation models assess:

    • Disease incidence and prevalence.

    • Social distancing impact.

    • COVID-19 related mortality.

Summary of COVID-19 Data Across Regions

  • Comparative analysis of cases per 100,000 residents across multiple European regions.

Current Healthcare Applications

  • Highlight examples:

    • Risks from aspirin usage in relation to COVID-19 outcomes.