Key Takeaways
- Most therapy discomfort isn’t a red flag. Feeling exposed, tired or irritable after a session is ordinary; feeling smaller as a person, session after session, is not.
- In a survey of more than 14,000 people in NHS therapy, around 1 in 20 reported lasting bad effects, and no single approach was more implicated than another.
- The strongest predictor of a bad experience wasn’t the method. It was not being given enough information about the therapy before it started.
- Some flags are bright lines rather than judgment calls: sexual contact, dual relationships, and breaches of confidentiality outside its stated limits.
- Raising a problem usually beats leaving quietly. Ruptures that get repaired predict better outcomes, not worse, with a pooled effect of r=.29.
You’re in the car afterwards, and you can’t tell whether that was a good session or a bad one. She interrupted you twice. She spent four minutes on a story about her own sister. When you said the thing you’d been building up to all week, she nodded and moved on. And now you’re sitting in a parking bay running the tape back, trying to work out whether you’re being oversensitive or whether something’s actually wrong here.
That’s the hard part about judging a therapist. The job involves being uncomfortable on purpose, so discomfort is a terrible signal on its own. What follows is the line the research draws between ordinary difficulty and a real problem, the flags that are bright lines rather than judgment calls, and the green flags worth naming, because most of this ends with you finding someone better rather than giving up.
Start with what isn’t a red flag
Plenty of things that feel wrong are just therapy working. Silence that runs longer than you’d like. Being asked a question you can’t answer. Leaving a session more stirred up than you arrived. Crying about something you thought you were over. Finding that the first three sessions felt like admin.
Early sessions are especially misleading, which is why knowing what a first session normally looks like is worth ten minutes before you write anyone off. Intakes are structurally awkward. Some excellent therapists have unremarkable ones.
Here’s a useful rule of thumb: discomfort about the material is expected. Discomfort about the person is data. Dreading the topic is normal. Dreading them is not.
The finding that reframes the whole question
Crawford and colleagues surveyed nearly 15,000 people receiving or recently finished with NHS outpatient therapy for depression or anxiety. Of the 14,270 who answered the question, 763 agreed that their therapy had produced lasting bad effects. That’s roughly 1 in 20, and the researchers called it a substantial minority (Crawford et al., 2016).
The interesting part is what predicted it. No particular type of therapy was more often associated with bad outcomes. What did stand out were people who said they hadn’t been given sufficient information about the therapy before it started, and people who weren’t sure what kind of therapy they’d been receiving at all. Participants from ethnic and sexual minorities were also more likely to report lasting harm, which the authors read as a case for stronger cultural competence in training.
That points somewhere specific. The most reliable early warning sign isn’t a dramatic personality. It’s fog. If several sessions in you couldn’t explain what your therapist is doing or why, that’s the flag the data supports.
Red flags that are bright lines
A handful of things aren’t judgment calls at all. The APA’s Ethical Principles of Psychologists and Code of Conduct sets them out, and equivalent codes govern counsellors, social workers and therapists in most jurisdictions.
Sexual contact. Standard 10.05 prohibits sexual intimacies with current clients outright. Standard 10.08 prohibits them with former clients for at least two years after therapy ends, and even after that only in the most unusual circumstances. There is no version of this that is a grey area.
Dual relationships. Standard 3.05 covers multiple relationships, where the therapist is also your employer, your friend, your business partner, or is in a relationship with someone close to you. Small towns and small communities make some overlap unavoidable, and the code recognizes that. What it doesn’t permit is a therapist creating an overlap that could impair their judgment or exploit you.
Confidentiality outside its stated limits. Confidentiality has real exceptions, typically risk of serious harm, certain court orders, and mandated reporting. A therapist should tell you what those limits are at the start. Discussing your material outside those exceptions is a breach, not a style.
If any of these has happened, the question isn’t whether to switch. Licensing boards exist for exactly this, and a complaint is a legitimate option rather than an escalation.
Red flags in how the work is run
These are less clear-cut and more common. Any one of them can happen on a bad week; the flag is a stable pattern.
They talk more than you do, most weeks. Some self-disclosure is deliberate and useful. A running commentary about their own life is not.
They can’t tell you what you’re working on. By a few sessions in, a therapist should be able to name an approach and a rough direction. Vagueness here is the fog effect Crawford’s data flagged.
The frame keeps slipping. Chronic lateness, sessions that get cut short, texts at odd hours, repeatedly rescheduling you. The reliability of the container is part of the treatment.
They push disclosures you’ve said you’re not ready for. Good trauma work is paced. Pressure to produce the worst memory on someone else’s timetable is a competence problem.
They minimise or argue with what you bring. Being challenged is part of the job. Being told your experience didn’t feel the way you say it felt is not.
They can’t tolerate being questioned. Defensiveness when you raise a concern is one of the most predictive flags on this list, for reasons the next section makes clear.
The flag only you can see
Strip away everything observable and one question is left: over a run of sessions, are you becoming more yourself or less?
Feeling worse during a hard piece of work is expected. Feeling smaller as a person over months is different. Watch for the version of yourself that shows up in the room. If you’ve started performing insight, managing their reactions, or editing what you bring so the session goes smoothly, something has gone wrong with the relationship even if nothing on the observable list has happened.
This matters more than it sounds, because the working relationship is the most consistently evidenced ingredient in the whole field. Across 295 studies and more than 30,000 patients, the quality of the alliance correlated with outcome at r=.278 (Flückiger et al., 2018). The relationship isn’t the soft part of therapy. It’s the part with the numbers behind it.
Green flags, which are easier to miss
Red flags get all the attention, so here’s what good looks like when you’re in it.
They explain what they’re doing and why, without being asked twice. They set out the limits of confidentiality early. They check how the work is landing rather than assuming. They can say “I don’t know” and “that’s outside what I’m trained for, here’s who is.” They remember what you told them. They notice when something shifts in the room and name it. They can take a correction without either collapsing or defending. And when you disagree with an interpretation, they’re curious rather than insistent.
One more, which people rarely think to look for: a good therapist is comfortable talking about the ending. Not urgently, and not as a threat, but as a normal part of the work.
Before you leave, try raising it
This is the part most articles skip, and it’s the part with the best evidence behind it.
Almost nobody raises it. Blanchard and Farber surveyed 547 adults in therapy and found 93% had been dishonest with their therapist at some point. More than 70% had concealed feelings about the therapy itself, and the two most common concealments were pretending to like their therapist’s comments and pretending to find therapy effective (Blanchard & Farber, 2016). Not out of malice. Out of deference, embarrassment, and not wanting to damage the relationship.
Which is a shame, because raising it works. Eubanks and colleagues pooled 11 studies covering 1,314 patients and found that ruptures which get resolved are associated with better outcomes, at r=.29 (Eubanks et al., 2018). A repaired rupture doesn’t just recover lost ground. It tends to leave the work stronger than before it happened.
So the sentence is worth having ready: “Something’s been sitting badly with me since last week and I want to say it out loud.” How they respond to that is itself the most informative test on this page. A therapist who can hear it, stay curious, and not make you manage their feelings about it has just shown you something no checklist could.
And if it doesn’t get better, you’re allowed to leave
Leaving a therapist who isn’t right for you is not rude, dramatic, or a failure of commitment. It’s a normal move that a large number of people make.
The ethics code is on your side here in a way people rarely realize. Standard 10.10 requires psychologists to end therapy when it becomes reasonably clear that the client is no longer benefiting, and to offer pretermination discussion and alternative referrals. Continuing to see someone you’ve stopped benefiting from isn’t loyalty. It’s the thing the professional code specifically tells the professional not to do.
It also matters more than you’d think who you land with next. A systematic review found the individual therapist accounts for around 5% of the variation in outcomes on average, and that the effect is larger when clients arrive with more severe difficulties (Johns et al., 2019). The harder your situation, the more the specific person matters, and the less sense there is in staying with a poor fit out of politeness.
Practically: you don’t owe a resignation speech. “I’ve decided to try a different approach, thank you for your time” is complete. Ask for a referral if you want one, most therapists will give a good one. If you’re going back to the search, our guide to finding the right therapist covers how to filter faster the second time, and platform-based services are usually the quickest place to change providers without starting the whole process from scratch, which the BetterHelp and Talkspace comparison gets into.
The takeaway
Most bad therapy isn’t sinister. It’s a mismatch that nobody named, running on quietly because the client was too polite to say anything and the therapist never asked. That’s the pattern the concealment research keeps finding, and it’s fixable in a way that genuine misconduct is not.
So hold two things at once. Discomfort is not evidence of a problem, and your read on whether you’re shrinking is not oversensitivity. Between those, there’s a conversation worth having before there’s a decision worth making.
And if you do leave, you’re not back at the start. You know more about what you need than you did before this therapist, which is the one thing a bad fit reliably gives you.
Looking for a better fit?
Knowing what to avoid is half of it. Our complete guide to finding the right therapist covers credentials, directories and the questions worth asking in a consultation call, and what to expect in a first session will stop you misreading ordinary awkwardness as a bad sign. If cost is part of why switching feels hard, how much therapy costs in 2026 lays out the lower-cost routes. And if this has left you wondering about therapy itself rather than this therapist, what the outcome research actually shows is the honest version.
Frequently asked questions
What are the biggest red flags in a therapist?
The bright lines are sexual contact of any kind, dual relationships where your therapist is also your employer, friend or business partner, and disclosing your material outside the stated limits of confidentiality. Below those, the patterns worth taking seriously are a therapist who talks more than they listen most weeks, one who can’t tell you what approach you’re working in, chronic slippage in reliability, pressure toward disclosures you’ve said you’re not ready for, and defensiveness when you raise a concern.
Is feeling worse after therapy normal?
Often, yes. Working on painful material stirs it up, and leaving a session more raw than you arrived is a common experience rather than a warning sign. The distinction worth tracking is between feeling worse about the topic and feeling worse about yourself. Temporary difficulty around what you’re processing is expected. A steady sense of shrinking across months is not, and it’s worth naming directly.
How do I know if my therapist is any good?
Watch for whether they explain what they’re doing and why, set out the limits of confidentiality early, check how the work is landing, can say “I don’t know” or refer you elsewhere, and can take a correction without defending. Beyond that, the most evidenced marker is the working relationship itself, which correlates with outcome at around r=.28 across 295 studies. If you can be straight with them and they can hear it, most of the rest tends to follow.
Should I tell my therapist I’m unhappy with therapy?
Yes, and the research is unusually clear on this. In one survey of 547 clients, more than 70% had concealed how they felt about the therapy, most commonly by pretending to like their therapist’s comments or pretending to find therapy effective. Meanwhile, ruptures that get resolved are associated with better outcomes, not worse. Raising it is both rarer and more useful than most people assume, and how your therapist handles it tells you a great deal.
How do I switch therapists without it being awkward?
You don’t owe an explanation or a farewell speech. “I’ve decided to try a different approach, thank you for your time” is a complete message, by email if that’s easier. Asking for a referral is normal and most therapists will give a considered one. Professional ethics codes actually require practitioners to end therapy when a client is no longer benefiting and to suggest alternatives, so leaving a poor fit is the outcome the code anticipates rather than a breach of it.
Sources
- Crawford MJ, Thana L, Farquharson L, Palmer L, Hancock E, Bassett P, Clarke J, Parry GD. (2016). Patient experience of negative effects of psychological treatment: results of a national survey. British Journal of Psychiatry, 208(3), 260-265. (Of 14,270 respondents, 763 reported lasting bad effects; no approach was more implicated, but insufficient information before starting and uncertainty about the type of therapy received both predicted harm.)
- Blanchard M, Farber BA. (2016). Lying in psychotherapy: Why and what clients don’t tell their therapist about therapy and their relationship. Counselling Psychology Quarterly, 29(1), 90-112. (547 adults; 93% reported some dishonesty, and over 70% concealed feelings about the therapy itself, most often by pretending to like a comment or to find therapy effective.)
- Eubanks CF, Muran JC, Safran JD. (2018). Alliance rupture repair: A meta-analysis. Psychotherapy, 55(4), 508-519. (11 studies, 1,314 patients: resolved ruptures were associated with better outcomes at r=.29.)
- American Psychological Association. Ethical Principles of Psychologists and Code of Conduct. (Standard 3.05 on multiple relationships; 10.05 and 10.08 on sexual intimacies with current and former clients; 10.10 on terminating therapy when a client is not benefiting.)
- Flückiger C, Del Re AC, Wampold BE, Horvath AO. (2018). The alliance in adult psychotherapy: A meta-analytic synthesis. Psychotherapy, 55(4), 316-340. (295 independent studies, more than 30,000 patients: alliance-outcome correlation r=.278.)
- Johns RG, Barkham M, Kellett S, Saxon D. (2019). A systematic review of therapist effects: A critical narrative update and refinement to Baldwin and Imel’s review. Clinical Psychology Review, 67, 78-93. (Weighted average therapist effect of 5%, larger where client severity was higher.)
Disclaimer: This article is for educational and informational purposes only and is not intended as a substitute for professional medical or mental health advice, diagnosis, or treatment. Always seek the guidance of a qualified mental health professional for any questions about your mental health or a medical condition.
If you are in crisis: In the US, call or text 988 (Suicide & Crisis Lifeline). In the UK, call 116 123 (Samaritans). Elsewhere, findahelpline.com lists country-specific services.
