From MRI to today: the core principles of strategic therapy you can use tomorrow

In the mid-1960s a small group at the Mental Research Institute in Palo Alto set out to test something unfashionable: that a problem could be resolved in a handful of sessions without anyone ever addressing where it came from.

What the 1974 paper claimed

The founding document is Weakland, Fisch, Watzlawick and Bodin’s paper in Family Process, and it repays reading for what it declines to say. It reports significant improvement in roughly three-quarters of 97 varied cases treated in ten sessions or fewer. It claims no cure for any specific condition. What it offers instead is a general logic, that ordinary difficulties become problems when mishandled by solutions that then take on a life of their own, plus procedures for interrupting the mishandling.

The theoretical scaffolding was already standing. Watzlawick, Beavin and Jackson’s Pragmatics of Human Communication (1967) had reframed relationships as communication systems governed by recurring patterned sequences instead of individual pathology. That reframe is why attempted solutions are still where a strategic formulation begins half a century later.

What changed between Palo Alto and Arezzo

Nardone studied at the MRI with Watzlawick as his mentor and returned to Italy to a caseload dominated by phobic and obsessive presentations rather than the family and psychotic work he had seen in California. What he built in Arezzo is more standardised than the original: named protocols for panic and phobic disorders, obsessive-compulsive disorder and eating disorders, each with its characteristic attempted solutions and a planned sequence of moves. The first session became a designed object, often carrying a prescription before the client leaves.

The two versions feel different in the room. The MRI approach reads as disciplined improvisation guided by a few firm principles. Nardone’s reads as a set of manuals under continuous revision.

How well do the numbers hold up?

Here the honest answer is more complicated than either camp usually admits.

The MRI’s own 1992 re-analysis of the Brief Therapy Center archive covered 285 cases with usable one-year follow-up and found roughly 44% full success and 24% partial success, averaging six sessions. Consistent with 1974, which is reassuring. It is also one clinical team auditing its own files.

The figures quoted for the Arezzo model are larger and rest on the same kind of footing. Bartoli and De la Cruz Gil (2023) summarise a longitudinal study covering 3,640 cases over ten years at an average of seven sessions, with reported efficacy of 95% for phobic and anxious disorders and 89% for obsessive-compulsive disorders. Those numbers circulate widely and are almost never accompanied by the caveat they need: they come from the centre’s own archive, without randomisation, blinding or independent outcome assessment. They describe what a practice recorded, not what a trial demonstrated.

Something closer to independent data has started to appear. A practice research network spanning several countries reported on 1,150 naturalistic cases, with complete symptom resolution in around 80% at an average of 5.4 sessions over 5.3 months (Vitry et al., 2021). Still naturalistic, still without a control group, but collected across multiple sites rather than one.

What to carry forward

Three things from 1974 survive intact. Treat the present maintaining pattern as the target. Assume the client’s own solution is doing more of the maintaining than the original difficulty. Keep the intervention brief enough that its failure becomes obvious quickly instead of after eight months of drift.

Everything else, from specific reframing techniques to the disorder protocols and the newer work on obsessive presentations, is refinement on that base. The disagreements with other schools begin there too, which is the subject of the comparison with CBT and solution-focused therapy.