Curious look into Observational Studies

Arete-Zoe – December 28, 2016 

Making sense of large data sets has always been a daunting task, and every tool that makes the life of an analyst easier is worth exploring. There are 45,000 observational studies on the database. This is how trial version of IBM Watson visualizes the data available. The dataset was minimally processed to reflect priority information requirements, in this case insight into the most commonly used study designs over time.

Observational study is a clinical study in which participants identified as belonging to study groups are assessed for biomedical or health outcomes. Participants may receive diagnostic, therapeutic, or other types of interventions, but the investigator does not assign participants to specific interventions.

Observational Study Design is described by MODEL (cohort; case-control; case-only; case-crossover; ecologic, family-based and other) and TIME PERSPECTIVE.

This slideshow requires JavaScript.

Observational Study Model: primary strategy for subject identification and follow-up:

  • Cohort: group of individuals with common characteristics who are examined or traced over a given time period.

  • Case-control: group of individuals with specific characteristics compared to group(s) with different characteristics, but otherwise similar.

  • Case-only: single group of individuals with specific characteristics.

  • Case-crossover: characteristics of case immediately prior to disease onset compared to characteristics of same case at a prior time.

  • Ecologic or community studies: geographically defined populations compared on a variety of environmental and/or global measures not reducible to individual level characteristics.

  • Family-based: studies conducted among family members, such as genetic studies within families or twin studies and studies of family environment.

Time Perspective – temporal relationship of observation period to time of subject enrollment.

  • Prospective: look forward using periodic observations collected predominantly following subject enrollment.

  • Retrospective: look back using observations collected predominantly prior to subject selection and enrollment.

  • Cross-sectional: observations or measurements made at a single point in time, usually at subject enrollment.

Curious look into Observational studies

Clarity on priority information requirements – as well as quality of data, is essential to get meaningful output from any analytical tool, no matter how useful and handy.