How we handle data

A short, plain-English summary for data-protection reviewers. This page describes how the product works. It is not legal advice and does not replace the agreement for an engagement.

What data we use

Public posts from public platforms and open forums, gathered for one research project at a time.
We do not collect from closed, login-gated or private health communities.

What we store, and what we minimise

We keep what the analysis needs: post text, source link, platform, date and engagement counts.
Once an author has been classified, profile details (bio, follower count, affiliations) are removed for anyone who is not a healthcare professional.
Professional details are kept for healthcare professionals, because they are the professional context of the comment.

What leaves the system

Exports are anonymised by default: author handles become [Patient], [Caregiver] or [HCP].
No profile links appear in deliverables unless the client, as controller, chooses otherwise. That choice is recorded in an audit log.
Patient and caregiver findings are reported in aggregate. We do not name, rank or profile individual patients or caregivers.

How healthcare professionals are treated

Only public professional commentary is in scope.
Naming healthcare professionals in a deliverable is an explicit opt-in by the client, and each export records that setting in the audit log.

Retention

Active projects keep their evidence.
When a project is archived, a purge date is scheduled. The default schedule is set in the platform and can be adjusted per engagement.
On purge, the raw posts are removed. Aggregate results such as barrier names, counts and frequencies remain, so reports still work.

Erasure

Erasure requests can be made by post URL, post ID or author handle.
Requests are honoured, including for healthcare professionals. We do not decline a request because a post was public.

Purpose limitation

Data serves one indication for one client.
There is no reuse across clients and no cross-indication database of people.
We do not train shared models on verbatim quotes.

Roles

Our standard position: the client is the controller and IndicationIQ is the processor.
A data processing agreement is provided as part of an engagement. Lawful basis and transparency duties toward the people in the data sit with the controller.

Sub-processors

The sub-processor register, including service, location and the data involved, is supplied with every agreement.

Questions from your data-protection team?

Get in touch and we will answer them before your first call. Our privacy notice covers the wider detail.