menu_open Columnists
We use cookies to provide some features and experiences in QOSHE

More information  .  Close

Health reform needs places where the system is allowed to learn

9 0
08.07.2026

(Version française disponible ici)

Every government that sets out to fix its health system eventually reaches for the same instrument. It centralizes. It consolidates authority, clarifies who is accountable and standardizes how things are done.  

The logic is sound and Quebec, like much of Canada, had real reasons to want a clearer line of sight from policy to performance. In 2023, the province passed a law that created Santé Québec, a Crown corporation that further centralized Quebec’s regional health networks, to oversee the day-to-day operations of the province’s health and social service institutions. The move was to make the system run more efficiently. 

But centralization solves one problem while quietly creating another. While it can make a system easier to run, it cannot, by itself, make the system better. These are different capabilities and confusing them is the most common way health reforms end up disappointing the people who were promised results. 

The remedy is not to abandon central stewardship. It is to pair stewardship with a small number of recognized innovation hubs: bounded, accountable, time-limited environments where new clinical, digital and organizational models are tested in real care settings before they are scaled. 

Testing care where it’s delivered 

A modern health system improves the way any complex organization does. It tries new approaches in real conditions to learn what works. It discards what doesn’t work and spreads successes. This requires somewhere to do the trying. A government that centralizes without deliberately protecting space for experimentation ends up with a system that is easier to direct but slower to improve.  

This matters because health care is a complex adaptive system. Interventions that look obvious on paper collide with workflows, incentives and human behaviour in ways no one can fully predict. A staffing model that succeeds in one hospital fails in another. An........

© IRPP - Policy Options