Remotely monitoring the health status of patients in all their activities of daily living is an interesting prospect. Yet an m-Health application devoted to patient monitoring should be able to deal with constrained and unstable operating conditions, such as limited power budget and disruption-prone network connectivity. This paper presents REGAS, a middleware system that makes it possible for a developer to define and enforce adaptive strategies, so their application can continuously adjust its behavior to changing operating conditions. Experimental results confirm that carefully designed strategies can have a significant impact on transmission delays, as well as on the power consumption of an m-Health monitoring application.