Why memory is a poor gauge of progress — and what regular measurement changes
1. The question at the start of the session
At the start of the session there is almost always the same question: “So, how have things been since last time?”
It is a good question. It is also the only one most treatments have at their disposal to find out whether they are working. And it has a design flaw worth knowing about: it asks a person to summarise two, four or six weeks of their life — in precisely the state they happen to be in today.
Someone having a bad day today remembers the past weeks as darker than they were. Someone having a good day underestimates how hard that Wednesday three weeks ago was. On top of that, we are poor at remembering trajectories. What sticks are the spikes and the ending — the worst moment and the last few days. The quiet, slowly rising middle part, which is exactly what matters in a therapy, leaves hardly a trace.
This is not a weakness of individual patients. It is the normal way memory works. It is just poorly suited as a measuring instrument.
2. What happens when you measure instead
Whether it helps to record the course of treatment regularly rather than ask about it has been well studied. In psychotherapy research this goes by the name of Routine Outcome Monitoring: patients are given a brief, standardised instrument at short intervals, and the result is fed back to the clinician.
The most comprehensive summary to date comes from Lambert, Whipple and Kleinstäuber (Psychotherapy, 2018). For one of the widely used systems, it pools 15 studies with 8,649 patients. Two of its results are noteworthy.
The first is sobering. Averaged across all patients, the effect is small (effect size 0.14). Anyone expecting measurement by itself to make therapy better will be disappointed. Where a course of treatment is going well anyway, feedback changes little.
The second is the real point. Among those patients whose course was unfavourable — “not on track”, in the technical term — the effect was markedly larger (0.33), and when the feedback came with concrete suggestions for how to proceed, it reached a medium magnitude (0.49). With feedback, this group improved with nearly twice the probability and deteriorated less often.
So the benefit does not lie in confirming success. It lies in noticing failure early enough.
Schematic illustration, not measured data.
And that is precisely what clinicians demonstrably struggle with. The same paper notes soberly that clinicians frequently fail to recognise deterioration and are poor at estimating how much a treatment will ultimately achieve. This is no reproach to their diligence. It is the predictable consequence of seeing a person for fifty minutes every two weeks and having to reconstruct the rest from their account.
3. What tracking actually delivers
From this evidence follow four things that regular recording can do and that conversation alone cannot.
It makes the direction visible. Individual values fluctuate; a line across twelve weeks shows a trend. Only the trend answers the question that matters.
It separates the bad day from the bad course. Whoever measures weekly can see whether a dip was an outlier or the beginning of a downward movement. In a conversation on a given day, the two cannot be told apart.
It sets a point in time for changing course. The value of measuring arises where a decision follows from the finding: adjust the dose, change the method, check for contributing factors. A trajectory nobody looks at is documentation, not steering.
It puts something in the patient’s hands. Whoever enters their own data over weeks sees their own curve — and with it the improvement that gets lost in everyday life, because we get used to the new state faster than we think.
The HBImed Therapy Tracking App is built for exactly this purpose: brief, regular entries by the patient, from which a trajectory emerges that the treating clinician has in summarised form before the next session. The evaluation is linked to the rest of the system, so that the trajectory sits alongside questionnaire results, test performance and brain function findings rather than apart from them.
4. And brain function?
The idea of not basing progress measurement on self-report alone suggests itself. A finding that can be measured can be repeated — a resting EEG takes a few minutes and asks nothing of the examined person that they would have to report. A progress assessment that places clinical impression, structured self-report and physiological measurement side by side would be more robust than any of the three on its own.
That is the direction we are working in. It is not yet where things stand. On the question of whether treatment-related changes are reliably reflected in electrophysiological indices — and from what magnitude a change is more than the normal fluctuation between two measurements — the data are thin. Anyone who today presents an EEG trajectory as proof of efficacy goes beyond what has been established.
5. Where measurement reaches its limits
Three limitations belong here, otherwise a useful tool turns into a promise.
What gets measured is what the instrument asks. A questionnaire captures what it was built for. What it does not contain — a side effect, a changed relationship, a new conflict — does not show up in the curve, however regularly one measures.
Numbers can crowd out the conversation. The benefit arises when the trajectory structures the conversation. It reverses when the curve replaces it.
Regular completion is a burden. Weekly entries cost patients attention, and they do so precisely when they are unwell. An instrument that is too long for this will not be completed — and a trajectory with gaps is more misleading than none at all, because it is exactly the bad weeks that are missing.
And the matter of course that is none: what people record about their wellbeing is among the most sensitive data there is. Anyone who collects such trajectories owes an account of where they are stored, who sees them, and what happens to free-text entries.
The honest summary is unspectacular. Regular measurement does not turn a bad therapy into a good one. It ensures that a therapy that is not working is recognised as such earlier — and, measured against what a lost phase of treatment costs, that is no small gain.
Reference
Lambert, M. J.; Whipple, J. L.; Kleinstäuber, M. Collecting and delivering progress feedback: A meta-analysis of routine outcome monitoring. Psychotherapy, 2018, 55(4), 520–537.