Therapeutic prescriptions that clients actually follow

A prescription is a good idea right up until the client leaves the room. What happens between sessions is where most strategic therapy succeeds or fails, and it turns less on the cleverness of the task than on a few unglamorous conditions.

Build it from the map, not from a technique library

“Stop checking the door” is advice. “Tonight, check the door once, then leave your keys somewhere you can see them from the bed, so the only way to check again is to get up and look” is a prescription assembled from this client’s ritual.

The second version is harder to forget and harder to quietly reinterpret, because she recognises her own evening in it. The mapping is what supplies that specificity.

Validation carries as much weight as logic

Duncan (1992) makes a point that repays attention. A prescription only carries weight if the client experiences it as validating what the problem is actually like for her, rather than as a theory imported from outside. A client who feels her suffering was taken seriously will attempt an odd-sounding task. A client who feels processed will not.

This is one reason the questions preceding a prescription matter as much as the prescription. They are what demonstrate that you understood the specific shape of this problem.

Written or spoken, and prescribing to more than one person

A task given only verbally is easier to half-remember and easier to soften into something more comfortable. Where the task is at all complex, have the client write the final version in her own words before she leaves. Copying your phrasing does not produce the same commitment.

Family prescriptions need an extra layer. Each person’s version should be specific to her own role in the pattern. A parent and child both told to “communicate better” will interpret it differently and blame each other when nothing shifts. A parent given one precise action and a child given a complementary one produces a cleaner test of whether the prescription did anything.

Concrete, small, checkable

Tasks that name a single action, a clear moment and an observable result get done. “Notice three times this week when you don’t check” is vaguer and more forgettable than “next time you feel the urge to check, set a two-minute timer before you allow yourself to look.”

Then ask, at the start of the next session, what actually happened. Not “how did it go”, but what she did, when, and what followed. That check is part of the intervention.

When a task fails, don’t start with the client

Stanton (1981) describes a family who agreed to tasks in session and then repeatedly failed to carry them through. His conclusion was not that the parents lacked motivation. It was that the therapist had been asking them to hold more responsibility than the family’s structure allowed, and that the resistant pattern was itself the most useful information in the case.

Non-completion is diagnostic. When a task does not happen, shrink it rather than repeating it more firmly, or, where the solution is unusually rigid, consider a paradoxical version of the same request.

When a task half-works

The most common result is neither compliance nor refusal. She did it twice out of seven days, or did a softened version of it.

Resist treating this as failure. Ask what was different about the two days it happened, which usually identifies the condition the task needs in order to run. Then build the next prescription around that condition rather than around the original task. A client who managed the delay on days she was not already exhausted has told you something specific about scheduling, and the useful next step follows from it.