You have had the same conversation at the kitchen table, on a walk and in the car outside the supermarket. Someone suggests therapy. “But will it actually help?” is a fair question when you have already spent so much energy trying.
Couples therapy can help with relationship distress, and research supports improvements in areas such as satisfaction and communication. It does not help every couple in the same way or guarantee staying together. The useful question is what the evidence suggests, where its limits are and how you will recognise worthwhile change in your own life.
What does the research show?
A 2020 meta-analysis brought together 58 studies representing 40 distinct samples and 2,092 couples. It found substantial average improvements in relationship satisfaction during therapy, alongside improvements in communication, emotional intimacy and partner behaviour. Couples on waiting lists did not show significant average improvement in satisfaction.
That is encouraging evidence, not a personal forecast. The analysis combined different study designs and therapies. Its samples consisted exclusively of opposite-sex couples, and the authors noted difficulty fully accounting for dependencies within studies. These limits matter when applying the findings to a particular couple.
An average improvement also cannot tell you whether a specific therapist, approach or appointment plan fits your concern. Use the research as a reason to explore qualified support, then ask how the proposed work relates to what you need.
Source:Roddy and colleagues, 2020: meta-analysis of couple therapy
Why one success percentage can mislead
Before trusting a headline number, ask what counted as success. Feeling more satisfied, showing a meaningful change on a questionnaire, communicating differently and remaining together are different outcomes. Ask who was studied and when the result was measured.
A five-year follow-up of a randomised trial involving 134 seriously distressed married couples illustrates this. Around half showed clinically significant improvement: 50% after integrative behavioural couple therapy and 45.9% after traditional behavioural couple therapy. Separately, 25.7% and 27.9%, respectively, had separated or divorced.
The trial compared two active therapies. Those figures do not tell us what would have happened to these couples without treatment, and they are not universal success or divorce rates. They show why a follow-up measure and a relationship-status measure should not be treated as interchangeable.
Nor should “success” become pressure to remain in a relationship. Your decisions, consent and safety still matter, whatever a study measured.
Source:Christensen and colleagues, 2010: five-year trial follow-up
What could useful change look like at home?
Start with a moment you want to handle differently. “We want to communicate better” is a reasonable hope, but it can be hard to review. “We want to discuss weekend childcare without insults or one person disappearing from the conversation” gives the work a clearer focus.
Here is an original example. Before seeking help, a couple repeatedly argues about visiting relatives. One person announces the plan; the other objects; both spend the evening defending their families. A possible goal is to discuss invitations before accepting and make room for both people’s limits.
At a review, they might notice that the last invitation still caused disappointment, but they discussed it before replying and neither person was punished for saying no. That would be a concrete change to explore, even though the wider disagreement remains.
These are examples of goals, not a clinical scoring system. Let the practitioner help you choose appropriate measures, and report differences in how each of you experiences the work.
Can support help before things become a crisis?
You can enquire before reaching a breaking point, but evidence for distressed couples should not automatically be applied to every happy relationship. Different preventive programmes need their own evaluation.
For example, a 2014 trial randomly assigned 215 married couples to the Marriage Checkup or a waiting list. The programme involved assessment and feedback visits, repeated a year later. Researchers found benefits in intimacy and some other relationship measures, with different patterns over follow-up. Participants were predominantly white and came from one northeastern US metropolitan area.
That supports exploring a specific structured check-up approach; it does not prove that every couple needs regular therapy or that any relationship app delivers the same benefit. Ask what the service actually involves and what evidence applies to it.
Choose a process you can understand and review
A recognised method name can help you ask informed questions, but it is not a guarantee. Ask the practitioner why their approach may fit your concern, what relevant training they have and what happens if the work is not helping.
Agree how often to meet and when to review progress, cost and fit. A research programme’s timetable is not a prescription for your relationship. There is no need to promise an unlimited series of sessions before understanding the plan.
Bring honest feedback about what happens between appointments. If an exercise causes distress or repeatedly turns into another argument, describe that rather than pretending it went well. The difficulty is information for the practitioner, not a homework failure.
BACP recommends checking qualifications and finding a therapist who feels appropriate for both people. Consent to attend needs to be voluntary. One person agreeing under pressure is different from both choosing to explore help.
Use a short review instead of asking only “Is it working?”
Before a planned review with your therapist, each write a few private notes if that feels safe. You do not need to agree on the answers first. This is an editorial preparation aid, not a treatment or diagnostic tool.
For example: “We argued twice this week, but we returned to the conversation without name-calling. I still dread bringing up money.” Both observations belong in the review. Improvement in one area does not erase distress in another.
- What concern brought me here, and is it still the main concern?
- What specific moment has changed, stayed the same or become harder?
- Do I understand the purpose of the work, and can I raise concerns about it?
- What do I want the practitioner to explain or reconsider next?
When things are not improving
Ask to discuss the lack of progress directly: “We are attending, but the same problem continues at home. How do you understand that, and what might need to change?” A review can consider the goals, approach, practical barriers or whether another kind of support is needed.
NIMH advises people to discuss concerns when treatment does not seem to help; sometimes a different provider or approach may be appropriate. That guidance is about reviewing professional care, not deciding alone that all therapy is useless after one uncomfortable conversation.
Do not turn a disappointing outcome into proof that either person did not try hard enough. Attending sessions cannot make someone accept responsibility, want the same future or freely choose the relationship. BACP is clear that counselling cannot simply fix a relationship or change a partner for you.
Source:NIMH: finding and reviewing professional care;BACP: couples counselling, expectations and limits
When joint therapy is not the next step
If there is abuse, coercion or fear of retaliation, a joint session is not automatically a safe place for honesty. The Hotline warns that information shared in couples therapy can be used against an abused partner. Confidential specialist support is the more appropriate starting point.
Where mental-health difficulties are part of the situation, discuss assessment and the suitable form of care with a qualified professional. Relationship exercises and apps do not provide that assessment. Your next step can be an enquiry, a review or individual support; it need not be a promise to stay together.
Source:The Hotline: why joint counselling can be unsafe with abuse
