Pharma intelligence platform step by step: how IndicationIQ runs an indication

This is how IndicationIQ runs a pharma intelligence project — on one indication, in one workspace, through to a deliverable your medical and compliance reviewers can open. Not a category lecture. The same four steps you will click through if you ask us to run your therapy area.

What you have when IndicationIQ is finished

A finished indication in IndicationIQ is not a sentiment timeline. You leave with:

• Ranked unmet needs in patient language, sized by frequency inside this indication — Patient reality, not a theme cloud.
• Misconceptions with the quote that shows the belief, ready for education briefs.
• HCP signals when the project is launch or medical communications — professional commentary, not a follower leaderboard.
• A quote bank: source, date, and verbatim behind every retained finding.
• A report you approve in-product, with handles anonymised to [Patient] / [Caregiver] unless you log an HCP naming exception.

That is the pack agencies drop into a workshop and in-house teams take to MLR. If you still own Meltwater or Brandwatch for brand monitoring, keep them. IndicationIQ is the layer that turns the conversation into named, evidenced recommendations.

Step 1 — You create an indication, not a campaign

In IndicationIQ the unit of work is one indication. You start the wizard: condition, therapeutic area, and a brief template that matches the job — Patient insight audit, Launch readiness scan, Misconception audit, or HCP voice scan. The template pre-sets phase, report sections, and the sources we usually switch on for that job. You can still edit every field.

Phase is a first-class choice

You pick patient support, HCP launch, or both before collection starts. IndicationIQ uses that choice to keep patient lived-experience findings and professional-capacity HCP commentary on different paths — different screens, different export annexes, different review rules. You do not sort that out in Excel after the fact.

Keywords the way patients and clinicians actually talk

The wizard asks for clinical terms, lay language, drug and regimen names, community phrases, and exclude terms. IndicationIQ can suggest a starting set from the condition. Brand-only keywords are how other tools miss the unbranded conversation — the six-month referral delay nobody tagged with your INN. We collect that conversation on purpose.

Geography you can defend

You choose global collection or named target markets. IndicationIQ keeps author geography separate from the market discussed in the post, and treats language as context — never as proof of country. Japanese text is not stamped “Japan.” English is not stamped “UK.” Ambiguous evidence stays unknown. Later, country scope on the readout lets you slice without pretending the model guessed a flag.

Public sources you can put in a DPA

You switch on Reddit, X, Bluesky, YouTube, open forums, and the other public connectors on the project. Closed or login-gated health communities stay off unless counsel and the controller have signed that on. The lookback is a decision — support-programme refresh versus a launch window — not an unbounded scrape “in case.”

Step 2 — IndicationIQ collects into your evidence feed

Collection is a button on the indication, not a separate vendor. IndicationIQ pulls from the sources you enabled, with the keyword and community set from step 1. Posts land in the evidence feed with platform, date, and URL. Content hashing stops the same thread inflating frequency later.

You can open the feed while it fills. That matters: you see what is in scope before extraction runs, and you can tighten keywords or sources if the first pass is noisy. Agencies use this to show a client the raw material without promising a finished narrative yet.

Public is not anonymous. IndicationIQ treats handles plus health discussion as personal data, scoped to this indication, erasable by URL, post ID, or handle. We do not reuse a corpus across clients or therapy areas. That is a product rule, not a policy PDF.

Step 3 — Extraction names the insight you will brief

This is the step listening dashboards leave to a week of analyst coding. IndicationIQ’s processing run proposes structured findings from the corpus and files them on the screens your team already recognises.

• Patient reality — named unmet needs in patient language, ranked by frequency in this indication.
• Misconceptions — the belief as posted, not a sanitised medical paraphrase.
• Journey — where in the pathway the conversation sits.
• HCP signals — professional-capacity commentary for launch and med-ed.
• Audience and language — how people say it, including non-English quotes kept verbatim.
• Quote bank — the trail behind every finding you keep.

Labels you can put in a workshop

IndicationIQ aims for names patients would recognise: “fear of the first self-injection at home,” not “injection anxiety.” Specificity is what lets a medical reviewer accept or challenge a row, and what lets an agency map a need to a programme module. Vague themes are how you end up re-coding the export by hand — we built the product so you do not.

Frequency is this project’s denominator

Counts are inside this indication, this phase, this window. They are not epidemiology and they are not global mention share. A need that recurs across authors is a design input. A need that appears once stays a signal, not a slide title. You see the count next to the finding, then open the quotes.

Patients stay aggregate; HCPs follow a different rule

IndicationIQ classifies author type as part of the run. Patient and caregiver surfaces stay aggregate in deliverables. Public professional-capacity HCP commentary can be named in the workspace, and in an export only when the controller opts in and we log it. There is no cross-indication person database. If that sentence is what your DPO needs to hear, it is already how the product is built.

Step 4 — You review, then IndicationIQ exports

Automation proposes. Your team still decides. In IndicationIQ you accept, merge, or drop findings against the verbatim trail. Pending review is visible. You do not ship “the model said so.” You ship a pack a human signed off.

1. Open a finding, read the quotes, keep or kill the label.
2. Keep patient and HCP sections on separate review paths — they already live on different insight screens.
3. Check country and language fields were not over-claimed.
4. Build the report in the workspace. Anonymisation is on by default.

What leaves IndicationIQ is the product: ranked needs, linked verbatims, phase-appropriate HCP signals, recommendations framed for the brief you picked in the wizard. Handles become role labels. Profile URLs stay out of the file. When the project ends you archive; we schedule raw-evidence purge and keep the aggregates so last year’s report still makes sense.

Need a read before a full indication?

Pulse is the same public-conversation idea on a shorter fuse: a bounded scan you can run when a bid or a scoping question cannot wait for a four-to-six-week project. It exports a directional PDF and can promote into a full indication, carrying the scope forward. Use Pulse to decide whether the conversation is rich enough. Use the indication loop when you need MLR-ready evidence.

See it on your therapy area

IndicationIQ is not a pharmacovigilance system and not a substitute for primary qualitative research when you need a designed sample. It is the intelligence workspace for teams whose bottleneck is turning public conversation into a deliverable reviewers will stand behind.

Ask for a demo on your indication. We will configure the condition, run the loop, and show you Patient reality, the quote bank, and the export — not a generic sentiment chart with your brand name typed into the search box.