Filter the evidence

System 3.0 offers a number of powerful filters that can help you find the most relevant relationships and pathways in any given system.

Filtering topics by role and type

For each topic, users can select to view the determinants (i.e., drivers), interventions, or outcomes associated with that topic.

Determinants, interventions, and outcomes are broken out into a set of four top-level type categories, while the determinants and outcomes have additional sub-categories.

Determinant types

High-level determinant and outcome types include behavioral, biological / health, environmental, and social, with sub-types under each of these. These types and sub-types are based on a combination of external, validated sources, including the Centers for Disease Control and Prevention (CDC), the Department of Health and Human Services (DHHS), the Agency for Healthcare Research and Quality (AHRQ), and the scientific literature.

Intervention types

High-level intervention types include diagnostic, therapeutic, preventative, and managing, and are based on the WHO's developing International Classification of Healthcare Interventions.

Filtering relationships

Every relationship in the System Graph can be filtered by the underlying evidence that supports that edge. This includes filtering edges by evidence quality, median effect size, sample size, and recency.

Filter	Definition
Evidence quality
Well-established	The number of sources supporting this relationship is within the top 10% of number of sources for all determinants in the disease system.
Significant	The number of significant findings supporting this relationship is within the top 10% of number of significant findings for all determinants in the disease system.
Highly Cited	The relationship is supported by at least one source that has a "cited by" count within the top 10% of all sources in the disease system.
Median effect size	The median effect size for all of the extracted statistical evidence for a given relationship (normalized to Cohen's d) is at least small (>0.2), at least moderate (>0.5) or large (>0.8).
Sample size	The relationship is supported by at least one study where the sample size of the study is at least 100, at least 1000, or at least 10,000.
Recency	The relationship is supported by at least one source that was published in the past year, or in the past five years.

Toggling between views

Each system can be viewed at one-degree upstream or two-degrees upstream of the topic, or one-degree or two-degrees downstream. One-degree means that all of the underlying evidence is looking at the direct association between two topics, while two-degrees shows how each of the topics are associated with a potential mediating topic.