Therapy app intake privacy risk begins at the point where someone is already vulnerable enough to ask for help. Before a first session, an online counseling service may ask for name, email, phone number, location, age, gender, relationship status, emergency contacts, payment details, insurance information, medication use, symptoms, crisis history, stressors, trauma, substance use, sleep problems, work or school pressure, and preferences for a therapist. Even if the user never completes signup, those answers can describe a mental health need more clearly than many medical bills do.
The FTC's BetterHelp action is the clearest warning for this category. The agency said the online counseling service shared consumers' sensitive health data, including information connected to intake questionnaires, with advertising platforms after making privacy promises. That does not mean every therapy app behaves the same way. It does mean users should treat mental health onboarding as sensitive from the first click, not only after a licensed clinician enters the picture.
The GoodRx case and the FTC's Health Breach Notification Rule materials reinforce the broader point: health-adjacent apps can sit outside the mental model many people have for doctor-patient privacy, yet still handle information that is deeply personal. HIPAA may apply in some relationships, but consumers should not assume every wellness quiz, counseling marketplace, chatbot, or symptom screener is protected like a hospital record. The privacy risk lives in the gap between how intimate the question feels and how commercial the surrounding data stack may be.
Intake quizzes can be especially revealing because they translate feelings into structured fields. A free-text note saying someone is afraid, grieving, depressed, drinking more, fighting with a spouse, or worried about self-harm can become a searchable support record. A checkbox can reveal panic attacks, fertility stress, LGBTQ identity, financial anxiety, job burnout, domestic conflict, or medication history. When those fields are joined with email, IP address, device ID, ad pixel, referral URL, or payment data, the user is no longer just browsing for care. They are creating a mental health profile.
The risk is not limited to advertising. Push notifications can reveal the app on a lock screen. Calendar integrations can expose appointment times. Insurance uploads can connect therapy interest to an employer-sponsored plan. Chat transcripts can sit in cloud backups. Downloaded invoices can remain on a shared family computer. A user who is trying to keep therapy private from a partner, parent, employer, roommate, or insurer can lose control through small operational leaks even when the core clinician relationship is respectful.
NIST's Privacy Framework and FTC guidance point toward a better design rule: minimize data before trust is earned. A counseling service should not ask for detailed symptom narratives until it can explain who sees them, how they are stored, which vendors process them, whether trackers fire on the page, and how the user can delete or export the account. Crisis resources should be easy to reach, but that does not justify broad analytics collection around the most intimate parts of the intake path.
A practical defense checklist is to slow down before typing the hardest facts into a commercial form. Verify the provider, read whether the service is a covered entity, use a private email address, turn off lock-screen previews, avoid signing in through social accounts, and decline optional marketing consent. If the app offers a quick matching quiz, answer only what is needed to match with a clinician and save the deeper story for a secure session. Ask how to delete intake data if you decide not to continue.
cloak's anti-exploitation frame matters here because a person looking for therapy is not a marketing segment. Active defense should warn when a mental health flow starts behaving like a retargeting funnel, when a form asks for more than matching requires, or when consent language buries advertising access inside a moment of distress. People should be able to seek help without turning their pain into a reusable identifier for platforms that have no role in care.