What Happens When Therapists Don't Track Progress

By Tanner Oliver, LCSW ·May 8, 2026

Therapists are trained observers of human behavior. They spend years learning to read facial expressions, notice patterns, interpret what's said and unsaid. It's reasonable to assume that this training makes them reliable judges of their own clients' progress.

The research says otherwise.

The prediction problem

In 2005, Hannan and colleagues conducted a straightforward study. They asked therapists to predict which of their clients would deteriorate — get measurably worse — during treatment. The therapists predicted a deterioration rate of about 1%.

The actual rate, measured by validated instruments, was 8%. The therapists missed the vast majority of clients who were getting worse.

This wasn't because they were bad therapists. It's because detecting gradual deterioration through session observation is genuinely difficult. A client who moves from a 12 to a 14 to a 16 on the PHQ-9 over six weeks doesn't look dramatically different in any single session. The change is incremental, masked by session-to-session variation in mood, topic, and presentation.

Without a number to anchor the comparison, the therapist relies on memory — and memory is subject to all the biases that make human judgment unreliable across domains. And importantly, the clients most at risk of deterioration are often the ones who present most normally in session — they've learned to mask, to put on a good face, to tell the therapist what they think the therapist wants to hear.

The overconfidence problem

The prediction problem is compounded by systematic overconfidence.

Walfish and colleagues surveyed over 600 mental health professionals in 2012. The results were striking. The average therapist rated their clinical skills at the 80th percentile — meaning they believed they were better than 80% of their peers. A full 25% placed themselves in the 90th percentile. Not a single respondent rated themselves below average.

This is mathematically impossible and psychologically unsurprising. Overconfidence in one's own abilities is one of the most robust findings in cognitive psychology. It appears in doctors, lawyers, investors, drivers, and every other profession where performance feedback is ambiguous or delayed.

In therapy, the feedback is especially ambiguous. Clients rarely tell their therapists they're not getting better — partly to avoid conflict, partly because they're not sure themselves, and partly because the therapeutic relationship creates a dynamic where positive feedback flows more easily than negative.

What this means in practice

When therapists don't track outcomes, several predictable things happen.

Deterioration goes undetected. The 5-10% of clients who get worse during treatment continue in an approach that isn't helping them — and may be actively harming them — because nobody has the data to see it.

Stagnation gets normalized. Clients who aren't improving but aren't getting worse occupy a gray zone. Without outcome data, the therapist can always tell themselves "these things take time." And sometimes that's true. But sometimes the client has been flat for six months, and continuing the same approach is a choice being made without information.

Treatment doesn't adapt. Evidence-based practice involves adjusting the treatment plan based on response. If a client isn't responding to the current approach, you try something different. But you can only make that adjustment if you know the client isn't responding. Without measurement, you don't know. You guess.

Termination gets delayed. When there's no objective measure of progress, there's no clear signal that treatment goals have been met. Therapy continues by default — sustained by the relationship, the routine, and the absence of a reason to stop rather than the presence of a reason to continue. This has real costs — financial, temporal, and emotional — for clients who have already achieved their treatment goals and would benefit from ending therapy and practicing their skills independently.

Conversely, premature termination also goes unrecognized. Without data showing that a client is only partially improved (say, a PHQ-9 of 11 that started at 20), the therapist may agree to end treatment when the client says they feel "much better" — even though objective measures would suggest the client is still in the moderate range and at risk for relapse.

The cognitive bias explanation

This isn't about therapist competence. It's about the cognitive biases that affect all human judgment.

Confirmation bias leads therapists to notice information that confirms their clinical hypotheses and discount information that contradicts them. If a therapist believes a client is improving, they'll notice the session where the client reports a good week and underweight the session where scores would have shown a spike.

The availability heuristic means therapists recall vivid moments — a breakthrough, a tearful disclosure, a moment of insight — more easily than the gradual trend of the data. These moments feel significant but don't necessarily correlate with measurable improvement.

Illusory correlation leads therapists to see patterns where none exist. If a therapist introduces a new technique and the client seems better the following week, the therapist attributes the improvement to the technique — even though the improvement might be regression to the mean.

These biases aren't character flaws. They're features of human cognition that affect everyone, in every domain. The solution isn't to be smarter. It's to use tools that compensate for the limitations of unaided judgment.

The corrective

Outcome measurement doesn't replace clinical judgment. It supplements it with information that clinical judgment reliably misses.

A therapist who reviews a client's PHQ-9 trend before each session starts with data. They know whether the client is improving, stable, or deteriorating before the conversation begins. They can use the session to explore what's behind the numbers rather than guessing whether progress is happening.

This isn't about reducing therapy to data. It's about ensuring that data has a seat at the table alongside intuition, experience, and relationship.

The therapists who track outcomes aren't the ones who distrust their clinical skills. They're the ones who understand that good judgment and good data are better together than either is alone.

Theracharts tracks client outcomes with 120+ validated assessments, trend charts, and clinical alerts — so you always know whether the work is working. Get started free.