Pick a condition. The map pulls the record live: the clinical
definition, which psychological and behavioral factors
the literature couples with it, what patients try, how they talk about
it to a search box and on social media, and the journey those pieces
sit on. Every number links back to the exact search that produced it.
What is what: bars measure lift, co-occurrence against the
PubMed-wide baseline, the method of the metabolite maps on this site.
Voices are live samples from public sources, matched by cue words. The
journey adapts the Corbin and Strauss illness-trajectory framework,
cited below. The original of this map drew on proprietary
clinical-outcomes data and industry digital-health research; this
public rebuild points the same frames at the open literature, where
every number can be checked. Literature attention is still not
prevalence, and a sample is not a population. The map is an initial
hypothesis to test against your own data; where a condition's own
record is thin, it draws on the adjacent area in MeSH so there is still
a signal to start from.
Around fifty live queries against PubMed, NLM MeSH, Wikimedia, Google suggest and public social feeds. A first visit also warms a month-long browser cache of PubMed-wide base rates, so allow up to a minute once; later runs are faster.
The condition
live · nlm mesh definition · pubmed record size
The patient journey
framework: illness trajectory, adapted from Corbin and Strauss · emotional state, motivators and voices inside each stage are live · the blank slots are the research plan
What the evidence couples with mental state
step 1 · live · lift over the PubMed-wide baseline · an initial hypothesis
Every count cites its source: PubMed rows link the exact query, top factors carry a 2005-2014 vs 2015-2024 trend, and a ✓ marks factors whose language also shows up right now in live search suggestions or the feed sample. Presence is a signal, not a count of evidence.
What patients try
live · therapies and self-treatment in the record, established or not
A mention is not an endorsement: these counts include studies of benefit, of harm, and of misuse alike. The self-medication and misinformation rows are where the record tracks what goes wrong.
Psychographic read
step 2 · labeled synthesis · archetypes assembled from the live signals on this page · segments, not people
Waiting for the live signals…
How patients talk about it
live samples · three channels, three registers
To the search box
google suggest · aggregated real queries
fetching…
On social media
public feed · recent posts
fetching…
To an LLM
openassistant open dataset · organic prompts
fetching…
Attention over time, by surface
live · each surface on its own honest window; where no public history exists, the panel says so
Research
pubmed abstracts per year · 2015-2024
Readers
fetching wikipedia pageviews…
Feed
posts per week, recent sample
Search
live suggestions · no public time series exists
…
LLM corpus
open prompt datasets · snapshot, not a series
…
How to interpret this. Read the map as an initial hypothesis to test with primary and secondary research. A long bar means the
research record couples this condition with that factor; publication
habits, funding and fashion all shape it. Small shares can still
matter clinically, and lived experience outruns the literature in
every direction. This is a tool for building patient-centered
products and services. It does not describe any individual and it
diagnoses nothing.